Plant phenotyping methods.

Root-TransUNet enables high-throughput phenotyping of Arabidopsis thaliana roots as a parameter in Heterodera schachtii parasitism →

Introduction Plant parasitism by sedentary plant-parasitic nematodes is a dynamic and continuously evolving process, accompanied by profound remodelling of host root system architecture across distinct infection stages. However, the physiology and anisotropic growth of Arabidopsis thaliana roots under Heterodera schachtii infection, together with complex lateral root proliferation and increasingly dense, overlapping morphology, pose substantial challenges for accurate image segmentation. Methods Here, we introduce Root-TransUNet, an optimised segmentation architecture that extends TransUNet by incorporating a composite loss function to improve boundary precision and structural continuity, as well as dual-stage strip-pooling (SP) modules to enhance elongated and directional root features. These adaptations address the unique morphological complexity of the infected root system. Additionally, we integrated Root-TransUNet into a high-throughput phenotyping pipeline and applied it to an existing dataset of ~120,000 images of 362 A. thaliana MAGIC recombinant inbred lines collected over several months of infection. By extracting root system architecture traits, including root surface area and estimated root volume across infection stages, we enabled stage-specific association analyses between host root growth and nematode performance across these genotypes. Results Root-TransUNet achieved strong segmentation performance, demonstrating improved structural continuity and boundary precision compared with widely used CNN- and Transformer-based baselines, including UNet++. Stage-specific analyses revealed that the relationship between host root traits and nematode performance changed as infection progressed. During establishment, nematode number was largely independent of initial root size and varied strongly among genotypes, whereas during the reproductive phase (10-30 dpi), greater root expansion coincided with reduced estimated nematode volume accumulation. Notably, nematode burden was largely independent of host root size before infection, indicating that root quantity was generally not a limiting factor for infection in this experiment. Discussion These results demonstrate that Root-TransUNet can robustly segment infected root systems across a wide range of nematode infection densities, providing a scalable image-analysis framework for studying plant-parasitic nematode parasitism in combination with host root phenotyping.

Why it matches plant phenotyping methods感染根系の画像セグメンテーション手法を開発し、高スループット表現型解析パイプラインに統合して根系形態形質を抽出しており、表現型取得・抽出法が中心的である。

abstractHere, we introduce Root-TransUNet, an optimised segmentation architecture that extends TransUNet by incorporating a composite loss function to improve boundary precision and structural continuity, as well as dual-stage strip-pooling (SP) modules to enhance elongated and directional root features.

A lightweight RPB-YOLO11-based detector improves mobile phenotyping of rice panicle blast →

Rice panicle blast detection is an important task in plant disease phenotyping. Field-based detection remains challenging because infected spike regions are often small, sparse, elongated, and affected by overlapping panicles, complex backgrounds, and variable illumination. In this study, we propose RPB-YOLO11, a lightweight YOLO11-based detector designed for rice panicle blast detection. The model uses a Lightweight Ghost Backbone (LGB) to reduce redundant computation. It uses Anisotropic Axial Stripe Attention (A2SA) to represent elongated panicle structures. It also uses Focal Multi-Scale Attention (FMSA) for multi-scale feature refinement and Adaptive Geometric Shape IoU (AGS-IoU) for geometry-aware localization. The model was trained and evaluated on a rice panicle image dataset containing 1,055 training images, 69 validation images, and 169 test images. On the test set, RPB-YOLO11 achieved 76.09% mAP50, 45.44% mAP50-95, 73.98% precision, and 72.75% recall with 6.21 GFLOPs. Compared with the YOLO11n baseline, it improved mAP50, mAP50-95, precision, and recall by 2.73, 2.12, 1.64, and 2.11 percentage points, respectively. An Android-oriented inference application supports local image inference, detection visualization, class counting, and diseased-panicle incidence estimation. These results suggest that RPB-YOLO11 provides a practical approach for image-based rice panicle blast survey.

Why it matches plant phenotyping methodsイネ穂いもちの画像検出モデルを開発・比較検証し、罹病穂率を推定する実用アプリまで構築しており、植物病害状態の画像ベース表現型取得が中心である。

abstractIn this study, we propose RPB-YOLO11, a lightweight YOLO11-based detector designed for rice panicle blast detection.

ECA-ModNet: a parameter-efficient network for unsound wheat kernel classification. →

Introduction Accurate classification of unsound wheat kernels is important for automated grain quality assessment, but improved recognition performance often comes at the cost of increased model complexity. Methods This study presents ECA-ModNet, a parameter-efficient convolutional network derived from EfficientNetV2-S. The architecture replaces two early-stage Fused-MBConv blocks with Mod-FusedMBConv blocks to introduce input-dependent local contextual modulation and replaces the squeeze-and-excitation modules in later stages with efficient channel attention to model local cross-channel interactions using fewer attention-related parameters. Experiments were conducted on the seven-class G600 wheat subset of the GrainSpace dataset. Results Across three independent runs, ECA-ModNet achieved an accuracy of 90.17 ± 0.21% and a macro-F1 score of 90.23 ± 0.21%, improving upon EfficientNetV2-S by 3.80 and 3.84 percentage points, respectively. The parameter count decreased from 20.19M to 16.49M, while FLOPs increased marginally from 2.90G to 2.95G. ECA-ModNet achieved accuracy statistically comparable to that of ConvNeXt-Tiny and InceptionNeXt-T while using substantially fewer parameters, and obtained 3.03-4.55 percentage points higher mean accuracy than six lightweight baselines. Discussion Ablation experiments identified two Stage 1 Mod-FusedMBConv blocks with a 3×3 context kernel as the configuration with the highest mean accuracy among those evaluated. These results indicate that ECA-ModNet offers a favorable accuracy-parameter trade-off for image-based classification of unsound wheat kernels.

Why it matches plant phenotyping methods小麦粒の状態（unsound kernel）を画像から分類するためのCNNを開発・比較・アブレーション評価しており、植物器官の状態推定手法が研究の中心である。

abstractThis study presents ECA-ModNet, a parameter-efficient convolutional network derived from EfficientNetV2-S.

BLAP: lesion-aware adaptive multi-scale visual prompt tuning for few-shot crop disease diagnosis. →

Introduction Applying general-purpose vision-language models (VLMs) to crop disease diagnosis presents three critical bottlenecks: reliance on large-scale annotated data, the high computational cost of full finetuning, and existing adaptation methods designed mainly for discriminative classification without sufficient visual-linguistic interaction for generative diagnosis. Methods We propose BLAP, an adaptive multi-scale visual prompt fine-tuning framework built upon BLIP-2. BLAP introduces an adaptive visual prompt fusion module (APFM) with learnable prompt vectors and a gating mechanism, together with a multi-scale pyramid feature fusion module (PFM). All BLIP-2 backbone parameters are frozen, and only 0.11% of the model parameters are optimized. Results On a few-shot dataset comprising 990 images from 11 crops and 33 disease categories, BLAP achieved 92.78% recognition accuracy, outperforming the BLIP-2+LoRA baseline by 21.67 percentage points. BLEU-4 and ROUGE-L scores reached 0.6507 and 0.7184, respectively, while inference latency increased by only 2.15%. Discussion BLAP provides a lightweight solution that balances accuracy, efficiency, and interpretability for crop disease diagnosis in resource-constrained settings. The proposed dynamic prompt fusion and multiscale pyramid adaptation strategy may also be extended to parameter-efficient fine-tuning of visionlanguage models in other domain-specific applications.

Why it matches plant phenotyping methods作物病害画像から病徴・病害状態を推定する視覚モデル適応手法BLAPの開発と評価が中心であり、植物病害フェノタイピング手法に該当する。

abstractWe propose BLAP, an adaptive multi-scale visual prompt fine-tuning framework built upon BLIP-2.

AI driven multi modal deep learning system for wheat disease detection, yield prediction, and crop health monitoring →

Sustainable wheat farming is challenging. Real-time information on crop health, disease transmission, and anticipated yields is essential for farmers. However, they frequently use slow, expensive, or non-communicative tools. This project develops a workable solution. There is no need for massive server farms because the entire system operates on a single graphics card. It incorporates images of wheat fields, Indian farming notes, greenhouse records, harvest statistics, and NASA meteorological data. Consider them as various “eyes” for crop photo analysis, and we tried several lightweight computer vision models. ConvNeXt-Tiny was slower but could operate on older equipment with 75% accuracy; EfficientNetB0 recognised wheat heads with 92% accuracy; and AgroMark, a hybrid solution that merged photo analysis with agricultural metadata (soil type, rainfall, increased to 87%, etc. Combining picture analysis with attention mechanisms (CBAM) allowed us to anticipate the amount of wheat that a field will yield based on these photo insights, and the results showed that our predictions were accurate, with an R 2 score of 0.97. Additionally, we developed a versatile detector that simultaneously detects disease, stress, head count, and pests. It is adjusted to deal with training data that is unbalanced (some diseases are common, while others are rare). As we packed everything into a 16-GB graphics card, we spent real time determining which strategies smaller training sets, removing weak features, and adjusting loss functions, work. We encounter real-world obstacles along the road, such as photographs from different locations not always match, mislabeled photographs from different locations not always match, mislabeled diseases, and neglected rare pests. Our step-by-step instructions, charts, and code are available.

Why it matches plant phenotyping methods小麦画像から病害・ストレス・穂数・収量などの植物形質・状態を推定するマルチモーダル手法を開発し、複数モデルの精度比較と実装上の検証を行っており、表現型取得・推定が研究の中心である。

abstractThis project develops a workable solution.

Backbone diversity beats text supervision: a systematic study of frozen multi-foundation model fusion for in-the-wild plant disease recognition. →

Automated plant disease recognition from field photographs remains challenging: models trained on laboratory datasets collapse on in-the-wild images, and the current state of the art on PlantWild-the largest open in-the-wild benchmark (18,542 images, 89 classes)-relies on text prototypes derived from a language model to reach 76.18% top 1 accuracy. We ask whether text supervision is truly necessary or whether the bottleneck is the diversity of the visual representation. Our central methodological finding is that backbone selection and fusion matter far more than classifier-head engineering : across 118 experiments, a simple fixed-weight linear-prototype combination on top of three complementary frozen backbones yields larger and more reproducible gains than any head-level adaptive routing mechanism we test. Specifically, through a systematic study of six frozen vision backbones (three CLIP, two DINOv3, and one DINOv2), three classifier heads, and four fusion configurations, completed in a single day on one consumer GPU, we establish three findings. (i) A single self-supervised backbone (DINOv2 ViT-L/14) already surpasses text-augmented MVPDR (77.56% vs. 76.18%). (ii) Concatenating three complementary backbones (DINOv2 + DINOv3 + CLIP) reaches 80.23% ± 0.41% (five seeds), exceeding the published MVPDR accuracy by +4.05 points and our own reproduction of MVPDR under an identical evaluation protocol by +7.96 points, without any language supervision. (The difference between the two deltas reflects evaluation-protocol differences-our split, model-selection criterion, and training schedule-rather than any discrepancy in the reported numbers; see Section 4.6.6 for a full reconciliation.) (iii) Every form of learned routing we test-per-class gating, backbone gating, sample-wise gating-is inessential; the gain is entirely attributable to backbone diversity and a simple linear-prototype scoring combination. On the smaller PlantDoc benchmark, the same principle transfers but with substantially higher seed variance: the best configuration reaches 80.09% at a favourable seed but 76.97% ± 1.48% over five seeds-a suggestive rather than robust gain. Beyond the accuracy headline, we provide a pathology-aware per-class analysis showing that DINOv2/v3 dominate on fine-texture lesion classes (rusts, mildews, and leaf spots) whilst CLIP's narrow advantage concentrates on organ/species-level identification (rice leaf and potato late blight). All primary claims are validated over five seeds, and all code, feature caches, and result files are released for full reproducibility.

Why it matches plant phenotyping methods植物病害の画像認識を対象に、複数の視覚基盤モデル融合と分類器を体系比較・検証しており、病害状態の推定手法が研究の中心である。

abstractOur central methodological finding is that backbone selection and fusion matter far more than classifier-head engineering

A distributed segmentation strategy developed for three-dimensional leaf trait quantification of tomato plants. →

Leaf parameters are crucial indicators reflecting the growing status of plants. Monitoring and analysis of leaf parameters significantly contributes to the improvement of crop yield and food quality. This study focused on three tomato plant varieties commonly grown in the Netherlands and proposed a fully automatic pipeline for leaf phenotyping. Three-dimensional (3D) point clouds of target plants were acquired with a specially designed imaging unit naming Maxi-Marvin. A semantic segmentation of plant organs was performed with PointNet++ model. To mitigate point cloud resolution decrement, the down-sampling operation in the baseline model was replaced with a distributed segmentation strategy. Leaf instances were further identified with Density-Based Spatial Clustering of Applications with Noise (DBSCAN), followed by a morphological phenotypic trait quantification based on 3D geometrical analysis. Target phenotypic traits including leaf length, leaf width, and leaf area. The evaluation results indicated that the distributed segmentation strategy achieved the best F 1 scores of 0.98 with block size set to 30,000. The Mean Average Errors (MAE) of leaf length, leaf width, and leaf area estimation were 2.09 cm, 1.78 cm and 8.98 cm 2 respectively. The estimation accuracies for leaf length, leaf width, and leaf area were 91.98%, 92.66%, and 89.67%, respectively.

Why it matches plant phenotyping methodsトマト葉の3D画像取得、器官セグメンテーション、葉インスタンス識別、形態形質推定を統合した自動フェノタイピング手法を開発・評価しており、方法が研究の中心です。

abstractproposed a fully automatic pipeline for leaf phenotyping

Application of the OliveID morphometric tool for the identification of archaeobotanical carbonized olive endocarps: evidence for morphological continuity with the modern Throumbolia cultivar. →

Introduction This study evaluates the applicability of digital morphometric analysis to images of archaeological carbonized olive endocarps as a proof-of-concept initial approach for archaeobotanical investigations. Conventional morphometric analyses of olive endocarps largely rely on manual measurements, limiting reproducibility and quantitative comparison. Methods Ten archaeological endocarps were selected from previously published archaeological assemblages based on the integrity of their outlines, apex-base morphology and overall preservation quality. Quantitative descriptors describing endocarp size, symmetry, curvature and contour geometry were extracted using the OliveID software and compared with a modern morphometric reference database comprising Greek and international olive cultivars. Results Reliable contour extraction and quantitative descriptor computation were successfully achieved for all archaeological specimens despite carbonization. Preliminary comparison of representative morphometric descriptors showed that the archaeological specimens were positioned within the morphometric variation observed among the modern reference collection. Hierarchical clustering consistently associated the archaeological endocarps with the modern Throumbolia morphotype, while distinguishing them from elongated, globular and mucro-bearing cultivars. Discussion These findings demonstrate the feasibility of applying digital image-based morphometric analysis to sufficiently preserved archaeological carbonized olive endocarps and indicate a similar morphometric affinity between the analyzed archaeological material and the modern Throumbolia cultivar. This proof-of-concept study highlights the potential of digital morphometric approaches for quantitative archaeobotanical investigations of archaeological olive remains, while emphasizing the need for larger archaeological datasets and standardized image acquisition to further validate the observed morphometric similarity.

Why it matches plant phenotyping methodsOliveIDを用いて炭化オリーブ内果皮の輪郭からサイズ、対称性、曲率、形状記述子を抽出し、デジタル画像形態計測の適用可能性と再現性を評価している。植物器官の形質抽出法が研究の中心である。

abstractThis study evaluates the applicability of digital morphometric analysis to images of archaeological carbonized olive endocarps as a proof-of-concept initial approach for archaeobotanical investigations.

Optimizing resource allocation in Miscanthus breeding via sparse testing designs for genomic prediction. →

Phenotyping high-biomass perennial crops is laborious and the rate of genetic gain in conventional perennial crop breeding programs is typically low. So, it is especially important to identify methods that produce efficiency gains in the breeding process. Miscanthus is a C4 perennial grass with favorable characteristics for producing biomass as a feedstock for biofuels and diverse bio-based products. Increasing biomass yield will increase profitability and environmental benefits, so it is a key target for Miscanthus breeding. In addition, the identification of well-adapted genotypes across a wide range of environmental conditions requires the establishment of multi-environment trials (METs). Sparse testing is a genomic prediction-based strategy that reduces the phenotyping costs in METs by selecting a subset of genotypes to evaluate in a subset of environments and then predicts the performance of the unobserved genotype-environment combinations. A Miscanthus sacchariflorus (MSA) population comprising 336 genotypes observed across three environments was analyzed implementing sparse testing designs. Three prediction models considering main effects (environments, genotypes, genomic) and interaction effects (genotype-by-environment; G×E interaction) were implemented for forecasting dry biomass yield (YDY), total culm (TCM), average internode length (AIL), and culm node number (CNN). Multiple calibration sets based on different compositions and sizes were considered to evaluate performance in terms of the predictive ability (PA) and the mean square error (MSE) for a fixed testing set size. The training set size ranged from 52 to 112 to predict a fixed set of 224 unobserved genotypes across all three environments. The results showed that the model accounting for G×E interaction consistently presented the highest PA and the lowest MSE: for CNN (PA: ~0.77, MSE: ~0.5) and YDY (PA: ~0.70, MSE: ~1.3) while for TCM and AIL these ranged from ~0.28 to 0.41 and ~1.3 to 4.3, respectively. Overall, varying training sets and allocation strategies did not affect PA and MSE, with 52 non-overlapping and 0 overlapping genotypes per environment as the optimal cost-effective allocation framework. This suggests that implementing sparse testing designs could significantly reduce phenotyping costs by fivefold, without compromising PA in breeding programs for perennial crops such as Miscanthus .

Why it matches plant phenotyping methodsスパーステスト設計とゲノム予測を用いて、複数環境での植物形質予測と表現型測定コスト削減を評価しており、表現型取得・予測手法が研究の中心である。

abstractSparse testing is a genomic prediction-based strategy that reduces the phenotyping costs in METs by selecting a subset of genotypes to evaluate in a subset of environments and then predicts the performance of the unobserved genotype-environment combinations.

AgriX-SENet: Squeeze-and-Excitation-based deep learning framework for explainable plant disease detection in sustainable agriculture →

Introduction Timely and accurate detection of plant diseases is essential for ensuring global food security and supporting sustainable agriculture. Conventional diagnostic approaches, such as manual inspection and laboratory testing, are often time-consuming, labor-intensive, and impractical for large-scale or remote agricultural environments. Although deep learning models, particularly Convolutional Neural Networks (CNNs), have significantly improved automated plant disease classification, they often lack interpretability and struggle to generalize under diverse field conditions. Methods This study proposes AgriX-SENet, an explainable deep learning framework that integrates Squeeze-and-Excitation (SE) blocks with a DenseNet121 backbone to enhance disease classification performance. The SE blocks recalibrate channel-wise feature responses to emphasize disease-relevant information while suppressing background noise. To improve model transparency, Grad-CAM, SHAP, and LIME were incorporated to provide visual and feature-level explanations of the model’s predictions. The framework was trained and evaluated using the Plant Pathology 2020 dataset containing four classes: healthy, rust, scab, and multiple diseases. Results AgriX-SENet achieved a training accuracy of 97.47% and a validation accuracy of 95.07%, outperforming fourteen state-of-the-art deep learning models. The classification report demonstrated high precision and recall across most disease categories, although the scab class exhibited comparatively lower recall, indicating an opportunity for further improvement. The explainability analyses consistently showed that the model focused on pathologically relevant regions of leaf images, validating the reliability of its predictions. Discussion The proposed AgriX-SENet framework effectively combines high classification performance with model interpretability, addressing a key limitation of existing CNN-based plant disease detection systems. Its ability to provide accurate and explainable predictions makes it a promising solution for scalable agricultural diagnostics. Future work will focus on improving classification performance for challenging disease categories and optimizing the framework for deployment on mobile and edge computing devices to enable real-time field applications.

Why it matches plant phenotyping methods葉画像から植物病害状態を推定する説明可能な深層学習フレームワークを開発・評価しており、植物表現型取得・判定手法が中心である。

abstractThis study proposes AgriX-SENet, an explainable deep learning framework that integrates Squeeze-and-Excitation (SE) blocks with a DenseNet121 backbone to enhance disease classification performance.

UNet-ECA-Bio: a biologically informed deep learning model for high-throughput micro-phenotyping of rice stem vascular bundles. →

Rice stem internal structure is a critical micro-phenotype influencing lodging resistance and yield; however, its analysis remains constrained by labor-intensive manual methods. Here, we present a publicly available dataset of 686 rice stem cross-sections, with 21,027 large vascular bundles (LVBs) and 19,342 small vascular bundles (SVBs) manually annotated. Five deep learning architectures were systematically evaluated, among which UNet-VGG16 achieved the best performance with a mean intersection over union (mIoU) of 87.4% (82.49% for LVBs and 74.41% for SVBs). An improved model, UNet-ECA-Bio, further raised mIoU to 89.32% and SVB IoU to 78.97% by integrating Efficient Channel Attention (ECA) and biologically informed class weighting using an image-level dataset. Leveraging these high-accuracy phenotypic predictions, genome-wide association studies (GWAS) indicated concordance between annotated and predicted traits, with SNP overlap rates of 96% (LVB count: 1,217/1,262), 43% (SVB count: 6/14), 98% (stem area: 122/124), and 100% (cavity area: 3/3) at −log10(p) ≥ 6. Meanwhile, compared with manual annotation (estimated 10–30 minutes per image), the proposed approach processed all 686 images within 10 minutes, representing a >600-fold increase in throughput. We further developed a user-friendly software tool, “Rice_Stem_Pre_V1.1.exe,” for automated phenotyping of 14 stem traits, providing a cost-effective platform for genetic studies of lodging resistance and yield improvement.

Why it matches plant phenotyping methodsイネ茎維管束の画像から複数の表現型形質を自動抽出する深層学習モデル、データセット、検証、ソフトウェアを中心的に開発しており、明確な植物フェノタイピング手法研究である。

abstractwe present a publicly available dataset of 686 rice stem cross-sections, with 21,027 large vascular bundles (LVBs) and 19,342 small vascular bundles (SVBs) manually annotated.

Anisotropic boundary-aware detection for cotton leaf diseases with boundary-decoupled regression and lightweight feature adaptation. →

Detecting cotton leaf diseases in open-field environments is challenging due to cluttered backgrounds, scale variation, and irregular lesion morphology. Conventional detectors rely on isotropic receptive fields and coupled box-regression losses, which limit their ability to localize elongated lesions with poorly defined boundaries. We present an anisotropic boundary-aware detection framework that propagates high-frequency boundary information across four successive pipeline stages. In the backbone, an Anisotropic Morphological Contrast Aggregation module (AMCA) enhances direction-aware representation and lesion-background contrast via re-parameterizable strip convolutions and high-frequency residual extraction. A Dynamic Semantic Boundary Transfer mechanism (DSBT) then captures boundary priors from shallow layers before they are lost to downsampling and injects them into the neck. A Morphological-Spectral Synergistic Feature Pyramid Network (MFS-FPN) preserves these cues during multi-scale fusion through spatial-domain operations compatible with edge hardware. Finally, an Anisotropic Boundary-Decoupled IoU loss (ABD-IoU) independently penalizes each of the four box boundaries and sustains optimization signals in high-IoU regimes via a logarithmic modulation factor. On the self-constructed Complex Cotton Leaf Disease dataset (CCLD; 6,856 images, 6 classes), the method achieves 78.50% mAP@50 and 65.00% mAP@50:95, improving the YOLOv11n baseline by 4.80% and 2.70% with only 2.73 M parameters at 202 FPS. Cross-domain evaluations on PlantDoc and RWD confirm consistent improvements. The framework runs in real time on NVIDIA Jetson edge platforms with INT8 quantization.

Why it matches plant phenotyping methods綿花葉の病斑・病害状態を画像から検出する手法を開発し、複数データセットとベースラインで性能検証しているため、植物表現型取得が中心である。

abstractWe present an anisotropic boundary-aware detection framework that propagates high-frequency boundary information across four successive pipeline stages.

What you plant may not be what you bought: morphological and genetic discordance in specialty Coffea arabica L. cultivars from Ecuador. →

The genetic identity of coffee cultivars is fundamental to the specialty coffee sector, where premium prices are paid under the assumption that the purchased planting material corresponds to the declared variety. However, many producing countries lack the certification infrastructure necessary to guarantee this identity in their informal seed systems, exposing producers to undetected varietal non-conformity. In this study, we examine a case from a specialty coffee ( Coffea arabica L.) farm in southern Ecuador where seeds labeled as Sidra (USD 100/kg) and Gesha (USD 500/kg) were purchased without genetic or phytosanitary certification. Using a combination of SSR-based DNA fingerprinting and quantitative morphological characterization, including plant architecture, leaf functional traits, and fruit characteristics, we documented varietal identity and assessed the discriminant capacity of morphological traits across the four resulting morphotypes. Using eleven microsatellite markers for SSR fingerprinting, we found that two of the four morphotypes did not match their declared commercial identity. One plant sold as Sidra was identified as compatible with Batian, a composite variety of Kenyan origin that is genetically unrelated to Ethiopian landraces. The plants acquired as Gesha corresponded to a pure Ethiopian landrace that is genetically similar to, but not identical to, the Panamanian Geisha reference accession T.02722. Only two morphotypes were confirmed as Sidra. Furthermore, the placement of Sidra within the Core Ethiopia genetic group is consistent with prior population-level analyses and with its likely status as a selected Ethiopian landrace rather than a variety of hybrid origin. Morphological linear discriminant analysis achieved 82.4% overall classification accuracy under leave-one-out cross-validation (LOOCV), with internode length dominating the first discriminant function (LD1 = 66.6%). These results demonstrate that varietal nonconformity in the specialty coffee seed sector can extend to the inadvertent introduction of genetically unrelated material and underscore the urgent need for accessible seed certification.

Why it matches plant phenotyping methodsコーヒー品種識別のための形態形質測定と判別分析が研究の中心であり、形態形質の識別性能をLOOCVで検証しているため、植物フェノタイピング手法の適用・検証に該当する。

abstractquantitative morphological characterization, including plant architecture, leaf functional traits, and fruit characteristics

SPVD-field: a task-oriented multi-task visual dataset for sweet potato virus disease under real field conditions →

Sweet potato virus disease (SPVD) is one of the most destructive diseases affecting sweet potato production worldwide, causing severe yield losses and posing a significant threat to food security. Vision-based intelligent diagnosis has emerged as a promising solution for large-scale SPVD monitoring due to its low cost and scalability. However, existing publicly available datasets for SPVD are extremely limited and typically focus on a single task, such as disease classification or lesion segmentation, under constrained imaging conditions. This lack of comprehensive, task-oriented datasets significantly restricts the development, evaluation, and fair comparison of advanced computer vision methods for SPVD analysis. In this study, we present SPVD-Field, a task-oriented multi-task visual dataset suite composed of two independently collected sub-datasets optimized for different computer vision tasks. Rather than constructing a single homogeneous dataset, SPVD-Field is deliberately organized into two complementary task-oriented sub-datasets: SPVD-DET, designed for disease detection with bounding-box annotations, and SPVD-SEG, designed for fine-grained lesion segmentation with pixel-level masks. The two sub-datasets were independently collected using different acquisition protocols optimized for their respective tasks, while sharing a unified semantic definition of SPVD symptoms, crop growth stages, and field environments. SPVD-Field captures substantial real-world variability in imaging scale, viewpoint, illumination, background complexity, and symptom manifestation, reflecting the inherent challenges of fieldbased disease diagnosis. We provide detailed documentation of data acquisition, annotation strategies, and quality control procedures, along with baseline benchmark results for both detection and segmentation tasks to demonstrate the usability and difficulty of the dataset. By offering a structured dataset suite rather than a single-task collection, SPVD-Field aims to support diverse research directions, including detection, segmentation, multi-task learning, and disease severity analysis, and to facilitate reproducible and comparable research in SPVD-related plant phenotyping.

Why it matches plant phenotyping methodsサツマイモの病徴を対象とする画像データセットで、検出・病斑セグメンテーション、データ取得・アノテーション・品質管理、ベンチマークを中心的に提供しており、植物病害状態の画像フェノタイピング手法・データ基盤に該当する。

abstractIn this study, we present SPVD-Field, a task-oriented multi-task visual dataset suite composed of two independently collected sub-datasets optimized for different computer vision tasks.

ReLeaf-SAM: reliability-guided detail compensation for SAM-based plant disease segmentation →

Accurate lesion segmentation is essential for automated plant disease analysis in precision agriculture. Although the Segment Anything Model (SAM) exhibits strong generalization ability, its direct application to plant disease images in natural field environments remains challenging due to cluttered backgrounds, dense leaf veins, uneven illumination, and frequent occlusions. In particular, SAM mainly relies on global structural cues and is often insufficiently sensitive to subtle lesion textures and weak local details, which can result in missed small or early-stage lesions and inaccurate boundary delineation. To address these limitations, we enhance SAM with a disease-specific detail compensation module for plant disease lesion segmentation. A ResNet-50based branch is employed to extract fine-grained local texture features that are difficult for SAM to capture. These fine-grained features are fused with SAM encoder representations and then injected into the SAM decoder, enabling more accurate lesion prediction while preserving SAM’s strong global modeling capability. More importantly, we propose a reliability-guided variational fusion framework to further improve the interaction between heterogeneous features. Specifically, instead of conventional similarity or addition-based fusion, we introduce an uncertainty-aware variational fusion strategy that explicitly quantifies the confidence of each feature stream. An uncertainty encoder models feature distributions probabilistically, and a variational fusion module dynamically assigns higher weights to more reliable features while suppressing uncertain or interfering responses. In addition, Kullback-Leibler divergence regularization is introduced to stabilize cross-feature alignment and improve fusion robustness. Extensive experiments on PlantSeg, PlantDoc-Seg, and ATLDSD demonstrate that the proposed method outperforms state-of-theart approaches, achieving DSC scores of 81.05%, 91.12%, and 88.27%, respectively. The proposed method addresses SAM’s weakness in fine-grained disease feature extraction, accurately identifies early and small lesions, and delivers reliable segmentation for field plant disease automatic diagnosis.

Why it matches plant phenotyping methods植物病斑を対象とする画像セグメンテーション手法を開発し、複数データセットで性能検証しているため、病害状態のフェノタイピング手法が中心である。

abstractwe enhance SAM with a disease-specific detail compensation module for plant disease lesion segmentation.

ShuffleNetV2 SSM MLCA: a lightweight recognition network for wheat fungal diseases. →

Introduction Wheat is one of the most widely planted staple crops worldwide and underpins global food security. Fungal diseases severely threaten wheat growth and trigger massive yield losses during cultivation. Traditional manual diagnosis is time-consuming and highly subjective, while existing deep learning models often struggle to achieve high accuracy and robustness in complex field environments. Accurate identification of these fungal diseases is therefore vital to secure grain production. Methods This paper constructs a lightweight convolutional neural network named ShuffleNetV2_SSM_MLCA for wheat fungal disease classification. First, the original basic blocks of ShuffleNetV2 are substituted with SS-Conv-SSM modules to strengthen the extraction of fine-grained lesion features amid visually analogous fungal disease samples; half convolution is embedded to cut down model computational overhead. Second, a Mixed Local Channel Attention (MLCA) unit is attached to the convolution branch of each SS-Conv-SSM module, which adaptively highlights discriminative disease features and filters irrelevant background noise. Standard training configurations and five-fold cross-validation are adopted for fair model evaluation. Results Comparative experiments reveal that the presented network reaches a classification accuracy of 91.35%, which surpasses the original ShuffleNetV2 baseline by 1.16 percentage points. Controlled ablation tests verify the independent performance gain of each core component: the SS-Conv-SSM module raises overall accuracy by 0.89%, and the MLCA mechanism brings an extra 0.27% accuracy increment. Discussion The proposed ShuffleNetV2_SSM_MLCA architecture strikes a favorable trade-off between model lightweight property and classification performance. It delivers a low-computation, high-precision recognition scheme for wheat fungal diseases and lays a solid technical foundation for real-time disease monitoring in intelligent agricultural scenarios.

Why it matches plant phenotyping methods小麦葉片の病斑特徴を画像から抽出し、植物の真菌病状態を分類する軽量深層学習手法を開発・検証しており、病害表現型の取得・推定が研究の中心である。

abstractThis paper constructs a lightweight convolutional neural network named ShuffleNetV2_SSM_MLCA for wheat fungal disease classification.

Pre-symptomatic detection of wheat stem rust using hyperspectral imaging and deep learning. →

Introduction Wheat stem rust (Puccinia graminis f. sp. tritici) remains a major threat to wheat production worldwide. Detecting the disease at the pre-symptomatic stage is important for earlier warning and more timely management. Methods We evaluated hyperspectral imaging and deep learning for pre-symptomatic wheat stem rust detection using a time-series dataset collected at 4-9 days post inoculation (DPI 4-9). Seven representative deep learning models were compared across DPI stages. A weighted cross-entropy strategy was then applied to the three strongest models, and model interpretability was examined using input gradient analysis, SHAP attribution, and vegetation-index screening. Results The weighted optimization increased overall F1-scores by 10.0%-18.4%. At the pre-symptomatic stage, the best model achieved an F1-score of 0.94 at DPI 4 and 0.99 at DPI 5, enabling detection before visible symptom development at DPI 6-7. Across the interpretability analyses, the 480-550 nm blue-green region emerged as the main source of information for pre-symptomatic detection, whereas the 750-870 nm near-infrared region contributed more general information on disease presence. Discussion These results show that hyperspectral imaging paired with deep learning can support accurate pre-symptomatic detection of wheat stem rust under controlled experimental conditions and provide useful evidence for future field-scale studies of early disease warning.

Why it matches plant phenotyping methods小麦の病害状態をハイパースペクトル画像と深層学習で検出する方法が研究の中心であり、時系列評価・モデル比較・性能改善・解釈性分析を含むため、植物フェノタイピング手法として含める。

abstractWe evaluated hyperspectral imaging and deep learning for pre-symptomatic wheat stem rust detection using a time-series dataset collected at 4-9 days post inoculation (DPI 4-9).

CoNutriNet: a dual-branch architecture with DenseNet and graph-enhanced attention network for coffee nutrient deficiency classification. →

Introduction Nutrient deficiencies in coffee plants significantly impact bean quality and yield, making timely detection crucial for successful cultivation. Current assessment methods rely on manual inspection, which is labor-intensive and time-consuming, posing challenges for large-scale field management. This approach often results in inconsistent evaluations and delayed interventions. Methods This study presents CoNutriNet, an automated deep learning architecture that integrates DenseNet121 with a novel Graph-Enhanced Attention Feature Network (GEAFNet) for classifying nutrient deficiencies in coffee leaves. DenseNet121 provides deep hierarchical and regional feature representation, while GEAFNet captures local, fine-grained spatial features through Inception, Ghost, and Efficient Channel Attention (ECA) modules. Furthermore, a Graph Convolutional Network (GCN) is included to model spatial dependencies and structural variations between leaf regions. Feature representations from both pathways are concatenated and refined using a Coordinate Attention (CA) module to enhance discriminative capability. Results Evaluation on the CoLeaf dataset demonstrates that CoNutriNet achieves an accuracy of 94.5%. The integration of lightweight attention mechanisms, dense connectivity, and graph-based modeling improves both performance and computational efficiency. Conclusion These results indicate that CoNutriNet achieves and efficient performance in nutrient deficiency detection in coffee crops, highlighting its potential for deployment in agricultural environments to support precision farming and optimize yield.

Why it matches plant phenotyping methodsコーヒー葉の栄養欠乏という植物状態を画像から分類する深層学習手法を開発し、データセットで性能評価しており、表現型取得・推定が研究の中心である。

abstractThis study presents CoNutriNet, an automated deep learning architecture that integrates DenseNet121 with a novel Graph-Enhanced Attention Feature Network (GEAFNet) for classifying nutrient deficiencies in coffee leaves.

ZDAM: a new deep learning model for bean leaf disease diagnosis. →

Introduction Accurate disease diagnosis is crucial for enhancing agricultural productivity and reducing postharvest losses, directly impacting food quality and safety. Traditional detection methods often rely on extensive feature modeling and perform poorly in complex field environments. Methods This study proposes a deep learning model called ZDAM, based on an improved ZFNet integrated with a dual attention mechanism. The classical ZFNet is first optimized to improve feature extraction efficiency. A combined channel and spatial attention mechanism is then incorporated to refine feature representation for disease identification in key crops. Finally, a residual module is added to boost accuracy. Results Evaluated on a dataset of 11,903 bean leaf images covering healthy leaves and four disease types, including leaf mould, rust, mosaic, and white spot, the model achieves an average recognition accuracy of 99.02%, outperforming MobileMamba, Vision Transformer, and Chest- OMD. Discussion This approach offers a scalable solution for automated disease monitoring, supporting postharvest quality preservation and sustainable crop production.

Why it matches plant phenotyping methods豆葉の病害状態を画像から推定する深層学習モデルを開発し、複数モデルとの性能比較も行っており、植物フェノタイピング手法が研究の中心である。

abstractThis study proposes a deep learning model called ZDAM, based on an improved ZFNet integrated with a dual attention mechanism.

RubberFormer: a transformer-based detection benchmark for rubber tree powdery mildew. →

Introduction Rubber tree powdery mildew is a major foliar disease that threatens the yield and quality of natural rubber. Its lesions are typically small, irregular, and embedded in complex backgrounds, making accurate automated detection difficult. Methods To address this challenge, we propose RubberFormer, an end-to-end detection framework based on a refined Transformer architecture for detecting small powdery mildew lesions in complex scenarios. RubberFormer adopts MobileNetV4 as a lightweight backbone, introduces the Hierarchical Attention with Local-global Optimization (HALO) module for multiscale local-global feature fusion, incorporates the Unified Cross-Attention Network (UCAN) to enhance multidimensional feature interaction, and applies Normalized Wasserstein Distance (NWD) Loss to improve small-object localization. Results Extensive experiments were conducted on PM-Dataset-Plus, which contains 9,765 images, and PD-40, a large-scale plant disease dataset containing 80,369 images across 40 disease categories and 8 crops. RubberFormer achieved superior detection accuracy and generalization performance compared with existing methods, while maintaining computational efficiency suitable for practical agricultural monitoring. Discussion These results demonstrate that RubberFormer is effective for detecting small and irregular rubber tree powdery mildew lesions under complex conditions. The framework has practical value for rubber tree disease monitoring and provides a transferable design strategy for agricultural vision tasks involving small objects and complex backgrounds.

Why it matches plant phenotyping methodsゴム樹の病斑という植物の病害状態を画像から検出するTransformer手法を開発し、複数データセットで性能検証しており、植物表現型取得が中心である。

abstractwe propose RubberFormer, an end-to-end detection framework based on a refined Transformer architecture for detecting small powdery mildew lesions in complex scenarios.

Identification of candidate genes involved in root gall formation during early infection of Plasmodiophora brassicae in B.napus . →

Clubroot disease, caused by Plasmodiophora brassicae , is one of the major constraints in rapeseed production. Breeding disease-resistant cultivars is the best way to control this devastating disease. However, breeding reliable resistant germplasm and genes is limited. Inactivation of susceptible genes has been shown to be a new and effective strategy for developing resistant crops. Therefore, we aimed to screen key candidate susceptible genes in this study. Firstly, we established a stable, high-throughput visualization method for identifying gall formation at the early stage of P.brassicae infection. At 14 days post-inoculation (dpi), the earliest time point with a clear record of scorable root swelling, remarkable variations in the speed of gall formation were observed among 85 genotypes. Secondly, genome-wide association studies (GWAS) were performed to identify genes involved in gall development. Three and two consecutive significant peaks were detected at 14 and 21 dpi, respectively. Thirdly, comparative transcriptomic analysis was conducted between 2AF195 and 2AF058 at 7 and 14 dpi; these two materials exhibit contrasting speeds of gall development. Gene clustering analysis revealed two opposite expression patterns at 14 dpi. One pattern comprised 1,383 genes downregulated in 2AF195 but upregulated in 2AF058, which were significantly enriched in 10 KEGG pathways, including Environmental Information Processing and Plant-pathogen interaction, and involved core repressors JAZ8/10 in the jasmonic acid (JA) signaling pathway, as well as nucleotide-binding site (NBS) protein-encoding genes. The opposite pattern consisted of 79 genes upregulated in 2AF195 but downregulated in 2AF058, which were enriched in an additional 10 KEGG pathways, predominantly related to Carbohydrate Metabolism and the Ubiquitin System. These genes were functionally annotated mainly as pectin methylesterases, xyloglucan endotransglucosylase/hydrolases (XTHs), and lignin biosynthesis-related enzymes. These findings demonstrated that distinct regulatory networks exist in different susceptible rapeseed genotypes. Finally, through the combined analysis of haplotype and transcriptome data, we co-localized and identified the candidate gene BnaC08g46100D , a nodulin-related gene belonging to the MtN21 transporter family. These results provide a theoretical basis for developing novel disease-resistant materials by editing the key susceptibility genes involved in root gall formation. The candidate genes identified in this study are the most promising targets for this purpose.

Why it matches plant phenotyping methods根こぶ形成を高スループットに可視化・判定する方法の確立が明示され、感染植物の病徴を測定する手法として研究の主要な技術要素になっている。

abstractwe established a stable, high-throughput visualization method for identifying gall formation at the early stage of P.brassicae infection.

Estimation of SPAD values in litchi based on improved LSTM with fusion of IoT and multispectral image texture features. →

Litchi is an important economic fruit in southern China, and its precision management relies on the rapid and accurate estimation of the Soil and Plant Analyzer Development (SPAD) values in leaves. Addressing the limitations of existing SPAD detection methods, such as limited rapid coverage, inadequate modeling of dynamic environmental interference, and shallow fusion of multi-source data, this study constructed an Internet of Things (IoT) system to collect real-time environmental data from a litchi orchard, combined with unmanned aerial vehicle (UAV) multispectral imagery to obtain canopy vegetation index and texture features. A Long Short-Term Memory (LSTM) network model integrated with a feature level attention mechanism (MLSTM) was proposed to fuse IoT time-series data, vegetation index, and high dimensional texture features for dynamic SPAD value prediction. The results indicate that multi-source feature fusion significantly improves SPAD estimation accuracy. The MLSTM model achieved optimal performance under the all-features situation, with a coefficient of determination (R²) of 0.897 and a root mean square error (RMSE) of 2.638, outperforming other comparative models. The attention mechanism effectively enhanced the model's focus on key features, improving feature utilization efficiency and model interpretability. The multi-source data fusion method and MLSTM model proposed in this study enable high precision, dynamic estimation of SPAD values in litchi leaves, providing reliable data support for precision fertilization, stress diagnosis, and yield prediction in litchi orchards, as well as theoretical support for promoting the practical application of this technology in smart agriculture.

Why it matches plant phenotyping methodsIoT・UAVマルチスペクトル画像から葉のSPAD値を推定するデータ融合システムとMLSTMモデルを開発・評価しており、植物形質取得手法が研究の中心です。

abstractthis study constructed an Internet of Things (IoT) system to collect real-time environmental data from a litchi orchard, combined with unmanned aerial vehicle (UAV) multispectral imagery to obtain canopy vegetation index and texture features.

Quantifying the reliability gap in cross-domain plant disease classification: benchmarking the limited efficacy of standard mitigation techniques under controlled-to-field shift →

Introduction Confidence calibration, selective prediction, out-of-distribution scoring, and deep ensembles are mature techniques in machine learning, yet their efficacy under the severe domain shift encountered when plant disease classifiers move from controlled laboratory imagery to heterogeneous field photographs has not been systematically benchmarked. Methods Models trained on PlantVillage were evaluated on PlantDoc leaf-level crop images under a parent-image-aware split protocol, and a suite of standard mitigation techniques was applied to characterize the reliability gap. Analyses included temperature scaling and selective prediction for a fine-tuned ResNet-50, quantitative image-level shift analysis, Grad-CAM visualization, simple target-aware adaptation baselines, frozen-feature backbone comparisons, and ensemble baselines. Results In the primary case study, a fine-tuned ResNet-50 suffered a 67.7-percentage-point accuracy collapse upon cross-domain transfer, while mean predicted confidence remained at 79.76%. Post-hoc temperature scaling reduced calibrated ECE to 0.3645 but left selective risk at 80% coverage at 64.30%. Quantitative image-level shift analysis confirmed large-effect-size differences in saturation ( d = 3.90), border edge density ( d = 3.33), and foreground-occupancy proxy ( d = 2.48) between the two domains, while Grad-CAM visualizations showed that the model shifts attention from lesion-centered regions in PlantVillage to background-dominated areas in PlantDoc. Simple target-aware mitigations, including adaptive batch normalization and feature moment matching, improved accuracy from 0.321 to 0.343 and 0.366, respectively, whereas DANN-style adversarial adaptation degraded performance to 0.252. A frozen-feature backbone comparison across five backbones showed that, within the energy-scoring frozen-backbone comparison, DINOv2-S/14 achieved the highest unknown-detection AUROC (0.764) and the lowest selective risk at 80% coverage (0.520), with paired Wilcoxon tests confirming statistically significant accuracy and macro-F1 differences across backbones. Two ensemble baselines were evaluated: a warm-start end-to-end ResNet-50 ensemble reduced calibrated ECE to 0.063 but achieved only 0.666 AUROC, while a lightweight DINOv2 linear-probe ensemble achieved 0.779 AUROC after calibration but under limited epistemic diversity. Discussion Neither ensemble established deployment-grade reliability: the best selective risk at 80% coverage across all configurations remained above 0.51. The principal contribution is a reproducible, deployment-oriented reliability characterization showing that standard post-hoc and lightweight adaptation techniques reduce but do not eliminate the severe reliability gap under controlled-to-field transfer in agricultural computer vision.

Why it matches plant phenotyping methods植物病害画像分類の信頼性・ドメインシフト・校正・選択的予測を体系的にベンチマークしており、病害状態を画像から推定する方法の技術評価が中心である。

abstracttheir efficacy under the severe domain shift encountered when plant disease classifiers move from controlled laboratory imagery to heterogeneous field photographs has not been systematically benchmarked.

Hyperspectral estimation of leaf chlorophyll under small-sample conditions via spectral augmentation and weighted ensemble learning. →

Introduction Under small-sample conditions, hyperspectral leaf chlorophyll estimation is affected by high-dimensional collinearity, measurement noise, and cross-source acquisition discrepancies. Existing studies often treat training-distribution expansion and model-error complementarity separately. This study proposed a physically constrained composite spectral augmentation-weighted ensemble framework for reproducible small-sample chlorophyll estimation. Methods Using 1,113 valid spectrum-label pairs from the leaf subset of the GreenHySpectra dataset in the 400-1000 nm range, spectra and chlorophyll reference values were matched by sample identifiers and divided into training and validation sets. Low-magnitude Gaussian noise and smooth wavelength warping were applied only to the training set. XGBoost, partial least squares regression, and ridge regression were optimized with Optuna using a CMA-ES sampler, and ensemble weights were calibrated by Bayesian optimization. An independent external set of 90 tomato leaf samples was used to evaluate transferability. Results Composite augmentation improved model stability and reduced validation error relative to the non-augmented baseline. The weighted ensemble model achieved the best internal performance, with R² = 0.6392 and RMSE = 8.8883. On the external samples, the model achieved R² = 0.498 and RMSE = 9.801. Discussion The proposed workflow integrates physically plausible augmentation, heterogeneous learner complementarity, and independent external validation. The external results indicate partial cross-source transferability while highlighting distributional and measurement-chain discrepancies that still limit absolute generalization.

Why it matches plant phenotyping methods葉のクロロフィル量という植物形質をハイパースペクトルから推定する手法を開発し、外部データで転移性を検証しており、表現型取得・推定が研究の中心である。

titleHyperspectral estimation of leaf chlorophyll under small-sample conditions via spectral augmentation and weighted ensemble learning.

Image analysis optimisation for carotenoid and anthocyanin content prediction in carrots: addressing colour parameter multicollinearity and genotypic diversity. →

Introduction Colorimetric analysis of food using the CIELab/Ch colour space (i.e., from digital images of samples) is an accessible, non-destructive method for carotenoid and anthocyanin content prediction. Literature presents very well-fit, but rudimentary, models for pigment estimation (e.g., single/multiple linear regressions). However, standardised methods that statistically account for the high multicollinearity between CIELab/Ch colour parameters, varying light conditions and colour calibration, and samples with high genotypic variability are lacking. Methods An image analysis optimisation was developed for the prediction of carotenoid and anthocyanin content of 16 carrot genotypes of different colours. Samples were photographed under six light conditions with a digital camera and image colour was calibrated before analysis with the CIELab/Ch colour space. Total pigment contents and individual carotenoid contents were analysed chemically via spectrophotometry and high-performance liquid chromatography, respectively. Partial least squares (PLS) regressions were used to assess the colour-pigment relationships to correct for high multicollinearity amongst the independent variables (CIELab/Ch colour parameters). Results/discussion The PLS models achieved satisfactory accuracy for the prediction of total carotenoid content ( ca. R 2 = 0.77) and total anthocyanin content ( ca. R 2 = 0.81) under all light conditions. The two models are suggested as robust approaches to total pigment prediction with multi-dimensional colour spaces, varying light conditions, and for a sample group of high genotypic variability. The carrot samples proved to have very high genetic diversity within each cultivar, resulting in unsatisfactory models for prediction of individual carotenoids ( ca. R 2 = 0.45) under the default light condition. However, all the results can be used to expand databases (towards artificial intelligence) and aid breeding programmes in search for higher concentrations of these interesting antioxidants for human health.

Why it matches plant phenotyping methodsニンジン試料の画像色解析を最適化し、化学分析値を用いてカロテノイド・アントシアニン含量を予測する手法を開発・検証しており、植物形質取得が研究の中心である。

abstractThe PLS models achieved satisfactory accuracy for the prediction of total carotenoid content ( ca. R 2 = 0.77) and total anthocyanin content ( ca. R 2 = 0.81) under all light conditions.

Spatially resolved quantification of wheat kernel vitreousness using hyperspectral imaging and spectral unmixing. →

Introduction: Wheat kernel hardness, vitreousness, and creaseness are key determinants of milling performance, yet they reflect different physical scales of grain structure and are not necessarily coupled. Methods: We developed a digital phenotyping framework based on hyperspectral imaging and spectral unmixing to quantify these traits at both kernel and cultivar levels in a diverse panel of common wheat. Pixel-level spectral unmixing resolved glassy, intermediate, and mealy endosperm components within individual kernels, enabling vitreousness to be expressed as a continuous spatial index. Results: The hyperspectral-derived vitreousness index showed moderate associations with kernel protein content and the protein-to-starch ratio, consistent with variation in endosperm packing density, but weak relationships with kernel hardness and crease geometry. Kernel hardness, primarily determined by puroindoline genotype, showed limited association with bulk protein and starch composition. Crease geometry, quantified using composite indices from RGB images, captured macroscopic grain features largely independent of both hardness and vitreousness. Discussion: These results demonstrate that hardness, vitreousness, and creaseness represent complementary but largely independent dimensions of grain quality, corresponding to molecular-scale adhesion, mesoscale packing, and macroscopic geometry, respectively. The proposed framework provides a scalable, non-destructive approach for resolving intra-kernel heterogeneity, enabling improved digital phenotyping for wheat breeding and quality assessment.

Why it matches plant phenotyping methodsハイパースペクトル画像とスペクトルアンミキシングを用いて小麦粒の硝子質を定量するデジタル表現型解析フレームワークを開発しており、形質取得手法が中心的である。

abstractWe developed a digital phenotyping framework based on hyperspectral imaging and spectral unmixing to quantify these traits at both kernel and cultivar levels in a diverse panel of common wheat.

ConvGeM-next: a deep learning framework for plant disease detection →

Introduction Plant diseases pose a major challenge to sustainable agriculture, particularly in regions that heavily depend on farming. Early and accurate identification of plant diseases is crucial for ensuring food production and minimizing crop losses. The rapid advancement of deep learning, particularly in convolutional neural networks (CNNs), has significantly enhanced plant disease classification performance. However, many models often struggle to generalize effectively in real-world scenarios due to challenges such as low-intensity visuals, low contrast between the background and foreground of the suspected sample, noise, and chrominance variation. Methods To address the challenges mentioned above, we introduce ConvGem-NeXt, an end-to-end deep learning architecture specifically designed for fine-grained plant disease classification, built on the ConvNeXt baseline model featuring enhanced generalization capabilities. More precisely, our method incorporates a learnable Generalized Mean pooling layer and ReLU activation in the ConvNeXt model to enhance spatial feature representation, and a custom classifier head that integrates batch normalization, ReLU activation, and dropout to mitigate overfitting and improve classification accuracy. Results We tested the presented model on two large-scale and diverse databases, PlantVillage and the PlantDoc. The model achieved 99.65% accuracy on the PlantVillage dataset and 94.69% accuracy on the real-world PlantDoc dataset, demonstrating the efficacy of our method for reliably classifying plant diseases. Discussion This work contributes to the rapidly growing field of agricultural automation by providing a reliable framework for timely disease diagnosis and supporting the enhancement of crop productivity.

Why it matches plant phenotyping methods植物病害を画像から分類する深層学習アーキテクチャを開発し、複数データセットで性能検証しており、植物の病害状態の取得・推定が研究の中心である。

abstractwe introduce ConvGem-NeXt, an end-to-end deep learning architecture specifically designed for fine-grained plant disease classification

Synthetic plant disease image generation to improve segmentation tasks in low-resource settings →

Accurate plant disease segmentation is often constrained by the availability of large, finely annotated datasets, particularly for rare diseases. This work presents a synthetic data generation pipeline that combines 3D leaf modelling with diffusion-based disease synthesis to address this limitation. Procedurally-generated leaf geometries are built in the 3D modelling package Blender to provide exact ground-truth masks, after which style-transfer is applied using Stable Diffusion, fine-tuned with Low-Rank Adaptation (LoRA) and guided by ControlNet conditioning to both preserve leaf structure and enforce correct lesion placement. The approach is evaluated on apple leaf diseases using a deliberately restricted subset of the PlantVillage dataset, simulating a controlled low-data-resource environment. Downstream task effectiveness is measured through leaf disease segmentation. The results show that combining data from the pipeline with limited real data leads to consistent improvements in segmentation performance.

Why it matches plant phenotyping methods植物病斑の画像セグメンテーション性能向上を目的に、3D葉モデルと拡散モデルによる合成データ生成パイプラインを開発・評価しており、植物病害状態の画像ベース推定が中心である。

abstractThis work presents a synthetic data generation pipeline that combines 3D leaf modelling with diffusion-based disease synthesis to address this limitation.

Perceptual graph kernels for image-derived plant trait interaction analysis in precision agriculture →

Latest imaging technologies play a vital role in the extraction of plant phenotypic traits in high ranges. Most existing analytical methods treat these traits as independent features, overlooking the complex interaction patterns that focus on plant responses to environmental stress. Proposed Perceptual Graph Kernel (PGK) framework model address the limitation in terms of image-derived phenotypic traits plat information graph structured interaction networks leverages perceptual similarity learning to capture higher-order phenotypic patterns. In the PGK framework, traits extracted from RGB (Red, Green, Blue) and multispectral imagery are encoded as nodes, and biologically meaningful relationships amongst trait pairs are represented as weighted edges. Extracted trait values are continuously transformed into perceptual states to enhance biological interpretability, and a graph kernel is employed to measure similarity between trait graphs. Experiments performed in an agricultural field with a precision agriculture dataset for plant stress phenotyping demonstrated that the proposed PGK achieved 93.8% classification accuracy, improving performance by 5.3 percentage points over the CNN baseline. The outcome results clearly highlight the effectiveness of the perceptual graph model for plant phenotyping and provide a robust, interpretable computational framework for sustainable crop monitoring and decision-support in precision agriculture.

Why it matches plant phenotyping methods画像由来の植物形質を抽出・関係グラフ化し、ストレス表現型分類を行う計算手法が研究の中心であるため。

abstractProposed Perceptual Graph Kernel (PGK) framework model address the limitation in terms of image-derived phenotypic traits

Multi-FusNet-convolutional neural network with improved Huber loss function for plant leaf disease detection and classification. →

Background Recently, plant disease detection and classification have become major concerns in agriculture. Early detection of plant diseases supports farmers to take precautionary actions to prevent the spread of infections across different parts of the plant. However, detecting and classifying plant leaf diseases remain challenging tasks due to the overlapping characteristics of different diseases. Methods To mitigate these limitations, this research developed a Multi-FusNet–convolutional neural network (Multi-FusNet–CNN) with an improved Huber loss function to classify multiple classes of plant leaf diseases. Here, a multipath residual network (Multi-RG) with cross-filtering fusion is integrated, and the pixel shuffling fusion method is developed for fusing low-level to up-sampled features. An improved Huber loss function is incorporated into the Multi-FusNet–CNN to effectively handle outliers and enhance the model’s generalization capability during training. Results The developed Multi-FusNet–CNN with improved Huber loss function achieved 99.95% accuracy, 99.13% F1-score, 99.87% recall, 99.27% precision, and 99.93% specificity, thereby outperforming existing conventional techniques. Conclusion The proposed Multi-FusNet–CNN model improved the generalization capability of the method during the training process on plant leaf disease detection and classification.

Why it matches plant phenotyping methods植物葉の病徴を画像から検出・分類するCNN手法の開発と性能評価が研究の中心であり、植物の病害状態を推定するフェノタイピング手法に該当する。

abstractthis research developed a Multi-FusNet–convolutional neural network (Multi-FusNet–CNN) with an improved Huber loss function to classify multiple classes of plant leaf diseases.

Comparative deep learning approaches for bean leaf disease recognition. →

Context Plant diseases are a serious danger to the world's food security since they drastically lower crop output. Traditional manual plant leaf inspection is time-consuming, labor-intensive, and frequently subjective. Recent developments in deep learning provide effective and scalable methods for image-based analysis-based automated plant disease identification. Techniques Three deep learning architectures-a proprietary Convolutional Neural Network (CNN), ResNet18, and Vision Transformer (ViT)-are used in this study to examine automated bean leaf disease identification. The Augmented iBean dataset, which has three classes-angular leaf spot, bean rust, and healthy leaves-was used to train and assess the models. Every model was trained using the same preprocessing and training settings to provide fair benchmarking. Receiver Operating Characteristic (ROC) curves, accuracy, precision, and confusion matrices were used to assess the model's performance. Outcomes ResNet18 fared better than CNN and Vision Transformer models, according to a comparative analysis. ResNet18 maintained a high level of computing efficiency while achieving 99% accuracy and 99.01% precision. Its better categorisation capacity across all disease categories was validated using confusion matrix and ROC analysis. In conclusion The study shows that ResNet18 offers the optimal trade-off between accuracy and efficiency and creates a standard benchmarking framework for bean leaf disease identification. The results demonstrate its applicability for real-time deployment in precision agricultural systems for better crop management and early disease identification.

Why it matches plant phenotyping methods豆葉の病徴を画像から認識する深層学習手法を比較・ベンチマークしており、植物病害状態の取得手法が研究の中心である。

abstractThree deep learning architectures-a proprietary Convolutional Neural Network (CNN), ResNet18, and Vision Transformer (ViT)-are used in this study to examine automated bean leaf disease identification.

Explainable deep learning-based comparative study for guava fruit and leaf disease classification: advancing agricultural diagnostics through AI. →

Introduction Early detection of plant diseases is essential for maintaining crop health and ensuring sustainable agricultural productivity. Guava fruit and leaf diseases, if not identified at an early stage, can lead to significant yield losses. Recent advances in deep learning offer promising solutions; however, challenges remain in achieving both high accuracy and model interpretability for practical agricultural deployment. Methods This study proposes an explainable deep learning-based framework for the classification of guava fruit and leaf diseases. A real-world dataset consisting of 527 annotated images across five classes-Disease Free, Phytophthora, Red Rust, Scab, and Styler and Root Rot-was utilized. Six hybrid model architectures were developed by integrating transfer learning backbones (VGG16, MobileNetV2, InceptionV3, and ResNet50) with custom convolutional neural network (CNN) classifiers. Model performance was evaluated using accuracy, precision, recall, F1-score, and class-wise metrics. To enhance transparency, Gradient-weighted Class Activation Mapping (Grad-CAM) was employed to visualize disease-relevant regions. Results Among all evaluated models, the proposed VGG16 + MobileNetV2 hybrid architecture achieved the best performance, attaining an accuracy of 96%, an F1-score of 0.96, and strong generalization across all disease classes. Comparative analyses using confusion matrices, ROC-AUC curves, precision-recall curves, and radar plots confirmed the superior and consistent performance of the proposed model over other hybrid configurations. Discussion The results demonstrate that combining deep feature extractors with lightweight architectures enhances both classification accuracy and computational efficiency. The integration of Grad-CAM provides meaningful visual explanations, increasing trust and interpretability in AI-assisted disease diagnosis. This framework shows strong potential for deployment in real-time smart farming systems and mobile-based diagnostic applications, particularly in resource-constrained agricultural environments.

Why it matches plant phenotyping methodsグアバの葉・果実画像から病害状態を推定する深層学習手法が研究の中心であり、複数モデルの比較評価とGrad-CAMによる説明可能性検証も行っているため、植物フェノタイピング方法論として採用する。

abstractThis study proposes an explainable deep learning-based framework for the classification of guava fruit and leaf diseases.

Evaluating UAV-based phenotyping strategies for Megathyrsus maximus . →

Effective high-throughput phenotyping is crucial for modern plant breeding, yet the optimal image acquisition parameters for UAV-based systems in forage crops remain poorly defined. We optimized UAV-based phenotyping methods for a Megathyrsus maximus biparental population, examining how ground sampling distance (GSD), environment, and harvest date affect the accuracy of RGB-derived digital traits in predicting yield and canopy height. Machine learning algorithms and mixed model analyses were applied to evaluate predictive power and heritability. Pixel count and Haralick's entropy showed strong correlations with conventional yield measurements, particularly in Environment 2, while most vegetative indices were poor predictors. Integrating machine learning substantially enhanced predictive power for green and dry matter yield (r > 0.80). For canopy height, machine learning models achieved correlations of 0.71 with ground truth measurements despite weak pairwise correlations. Mixed model analysis revealed high broad-sense heritability (0.7 < H 2 < 0.87) for yield traits, pixel count, and entropy, while vegetative indices and canopy height showed greater environmental susceptibility. Moderate GSD resolutions (0.5–1.0 cm) consistently outperformed both very high (0.27 cm) and very low (1.5 cm) resolutions. Coincidence index analysis demonstrated 80% correspondence between top genotypes ranked by pixel count and conventionally measured dry matter yield. This study provides an optimized framework for UAV-based phenotyping in M. maximus , demonstrating that combining advanced digital traits with machine learning accurately predicts key agronomic traits and significantly enhances genotype selection efficiency in forage breeding programs.

Why it matches plant phenotyping methodsUAV画像取得条件、RGBデジタル形質、機械学習による収量・草高推定を最適化・検証する研究であり、植物表現型取得法が中心的です。

abstractWe optimized UAV-based phenotyping methods for a Megathyrsus maximus biparental population, examining how ground sampling distance (GSD), environment, and harvest date affect the accuracy of RGB-derived digital traits in predicting yield and canopy height.

Digital morphological data can generate accurate pre-emergence herbicide dose-response curves in Chenopodium album L. →

Introduction Herbicide dose-response assays are routinely implemented to compare herbicide resistance among weed biotypes, which requires plant biomass to estimate the dose that reduces growth by 50% relative to untreated plants (GR 50 ). The Phenospex TraitFinder is a high-throughput, non-destructive, digital phenotyping system that collects data from 7 spectral parameters and 13 morphological parameters, including Digital Biomass (DB), which offers the opportunity for researchers to eliminate the time and labor associated with manual biomass collection. However, DB is the product of 3D Leaf Area and Plant Height (PH) Mean, making it a measurement of plant volume and an indirect indicator of biomass. While DB is highly correlated with true biomass, digitally collected plant volume data has not been implemented for dose-response assays or assessed for accuracy relative to true biomass data. Additionally, inaccurate PH measurements could impact the accuracy of DB measurements. Methods This study sought to assess the accuracy and utility of DB and the 19 remaining parameters in dose-response assays by comparing dose-response curves and GR 50 estimates generated from digital data and fresh biomass (FB) data. Accuracy of PH measurements were also assessed by comparing digital and manual measurements with the paired t-test. Pre-emergence dose-response assays using fomesafen and atrazine were implemented with common lambsquarters ( Chenopodium album L.). At 21 days after treatment, manual measurements of FB and PH were collected following digital data collection. Results Consistently strong correlations ( r = 0.97, P < 0.05) were observed between digitally collected data and their equivalent manual measurements. Comparisons of the dose-response curves indicated that only 3D Leaf Area, DB, Convex Hull Area, Projected Leaf Area, and Voxel Volume Total generated highly similar curves and GR 50 estimates relative to FB data, indicating that any one or all of these parameters could be utilized instead of FB. Small differences (approximately 1.06 to 1.77 mm) between manual and digital PH measurements were identified with the paired t-test, but since DB consistently produced similar dose-response curves and GR 50 estimates relative to FB, these differences did not impact the accuracy of DB measurements. Discussion Without requiring manual biomass collection, turnaround time for dose-response and other phenotyping assays decreases and allows faster sharing of research. Furthermore, herbicide-resistant plants can be preserved for phenotyping at later growth stages, tissue collection, and to produce progeny for future experiments.

Why it matches plant phenotyping methodsデジタル表現型システムで植物体積・草丈などを取得し、手作業の生体重測定との精度比較および除草剤用量反応曲線への有用性を検証しており、表現型取得法が中心です。

abstractThe Phenospex TraitFinder is a high-throughput, non-destructive, digital phenotyping system that collects data from 7 spectral parameters and 13 morphological parameters

Deep learning techniques for early detection and classification of leaf diseases in crops. →

Introduction Rapid population growth and climate change have intensified the need for sustainable agricultural productivity. Plant leaf diseases significantly impact the crop yield, quality, and food safety, necessitating accurate and automated detection methods. Methods This study proposes a deep learning (DL)-based framework for automated detection and classification of tomato and soybean leaf diseases. The proposed framework is trained and evaluated over a large-scale datasets comprising 16,012 tomato leaf images and 6,410 soybean leaf images. Multiple convolutional neural network (CNN) models, including DenseNet121, MobileNetV2, and InceptionV3, are employed for classification. Object detection is performed using YOLOv12. To enhance interpretability, Gradient-Weighted Class Activation Mapping (Grad-CAM) is integrated. Furthermore, a novel Hybrid Attention-Based Stacking Ensemble Model is developed using ResNet152V2, VGG19, and EfficientNetB0, combined with Convolution Block Attention Module (CBAM) and spatial attention mechanisms. Results The CNN models achieved classification accuracies of 97% for DenseNet121, 98% for MobileNetV2, and 99.94% for InceptionV3. YOLOv12 attained a mean average precision (mAP) of 99.5%. The proposed hybrid ensemble model achieved an accuracy of 99.18%, demonstrating improved feature learning through combined channel and spatial attention. Grad-CAM visualizations confirmed that the model effectively identifies the disease-relevant regions. Discussion The results indicate that the proposed framework has attained a high accuracy, robustness, and interpretability for plant disease detection. The integration of attention mechanisms and explainable AI enhances model reliability and transparency. This framework shows a strong potential for the real-time agricultural monitoring, although further validation across diverse crops and real-world field conditions is required.

Why it matches plant phenotyping methods植物葉の病害状態を画像から自動検出・分類する深層学習フレームワークの開発であり、病害表現型の取得・推定が研究の中心。

abstractThis study proposes a deep learning (DL)-based framework for automated detection and classification of tomato and soybean leaf diseases.

Distilled vision transformers with CNN fusion for robust cashew apple maturity prediction. →

Introduction Cashew apple is a nutrient-rich fruit containing abundant minerals, vitamins, and energy. However, its fleshy texture and delicate skin significantly limit its storage life and market value. Accurate maturity grading is therefore essential for improving post-harvest management and transportation efficiency. Methods This study proposes a lightweight vision transformer (ViT) student model trained using multi-granular knowledge distillation (KD) from a stronger data-efficient image transformer (DeiT)-Base teacher. The distillation framework integrates response-based soft-label supervision, attention transfer, and token-level feature regression to enhance representation learning under limited data conditions. Auxiliary lightweight architectures, including MobileNet, ConvNeXt, and EdgeNeXt, were trained independently to provide complementary predictions, and a weighted fusion strategy was employed for ensemble evaluation. Results The proposed ensemble ViT-KD with EdgeNeXt achieved 90% accuracy under the evaluated test split. To ensure statistical reliability and address potential partition bias, a stratified fivefold cross-validation was conducted on the dataset, yielding a mean accuracy of 86.89% ± 2.89% with consistent F1 scores and recall. The relatively low variance across the folds indicates stable internal generalization. Comparative experiments with conventional convolutional neural network (CNN) baselines and lightweight CNN baselines such as MobileViT-S and ShuffleNetV2 were performed, with the proposed ensemble framework achieving improved accuracy while maintaining computational efficiency. Computational analysis indicates that the stand-alone distilled ViT maintains a real-time inference capability of 8.79 ms per image, which supports suitability for edge-oriented agricultural applications. Discussion These results highlight the effectiveness of knowledge-distilled lightweight transformers for data-efficient maturity grading of cashew apples.

Why it matches plant phenotyping methodsカシューナッツ果実の成熟度という植物器官の状態を画像から推定する手法を開発し、交差検証・比較実験・推論速度評価まで行っており、フェノタイピング手法が中心である。

abstractThis study proposes a lightweight vision transformer (ViT) student model trained using multi-granular knowledge distillation (KD) from a stronger data-efficient image transformer (DeiT)-Base teacher.

From 3DGS scenes to plant traits: a scalable extraction and segmentation framework for muskmelon phenotyping →

Automated quantification of plant-level development from multi-plant greenhouse scenes requires separating individual plants from shared scene-level reconstructions and quantifying organ-level development, a challenge that single-plant acquisition workflows do not directly address. This study presents an end-to-end phenotyping pipeline built on 3D Gaussian Splatting (3DGS) and a post-reconstruction extraction framework, LCR-GS, designed to isolate plant instances from full greenhouse scenes without scene-specific model retraining. LCR-GS integrates zero-shot 2D cues with multi-view lifting, geometric clustering, and chromatic refinement to convert large scene-level reconstructions (~2M Gaussians) into compact per-plant subsets (~16K Gaussians). Experiments on greenhouse-grown muskmelon at the early vegetative stage demonstrate high plant-extraction precision (0.933) and strong organ-level instance segmentation (mean AP50 = 0.924). Plant height and leaf count are validated against manual measurements (height R² = 0.98, RMSE = 1.88 cm; leaf count R² = 0.86), whereas additional morphological traits, including leaf area, leaf area index, mean internode length, and stem node count, are reported as pipeline-derived descriptors for within-cohort comparison. By decoupling semantic inference from reconstruction, the pipeline reduces scene-scale data by over 99% and provides a practical route to derive compact per-plant 3D representations from multi-plant greenhouse imagery for downstream organ-level analysis.

Why it matches plant phenotyping methods3DGS画像から個体・器官を抽出し、植物形質を定量化するフェノタイピング手法の開発と検証が中心である。

abstractThis study presents an end-to-end phenotyping pipeline built on 3D Gaussian Splatting (3DGS) and a post-reconstruction extraction framework, LCR-GS, designed to isolate plant instances from full greenhouse scenes without scene-specific model retraining.

ReMA: a residual gated multi-head attention module for MobileViT in sugarcane diseases and disease recognition. →

Purpose/significance Sugarcane is a vital global crop, critical for sugar and energy production. The accurate and timely identification of its leaf diseases is paramount for sustaining the health and stability of the sugarcane industry. While deep learning models offer promising solutions, their deployment on mobile or edge devices is often hindered by substantial model size and high computational demands. Conversely, existing lightweight models frequently compromise on feature extraction capabilities and recognition accuracy. To bridge this gap, this study develops an architecturally improved lightweight model designed to achieve both high accuracy and computational efficiency. Methods We propose the ReMA-MobileViT model, which significantly enhances feature representation by incorporating a newly designed Residual Multi-head Attention (ReMA) module. This module ingeniously leverages a multi-head attention mechanism to capture richer contextual information from diverse subspaces, while its residual connection structure effectively mitigates network degradation and facilitates robust gradient flow. The proposed model underwent rigorous training and evaluation on a comprehensive Mendeley Data repository for classification tasks. Results Experimental evaluations demonstrate that the ReMA-MobileViT model achieves an outstanding classification accuracy of 99.02% on the sugarcane leaf disease dataset, substantially surpassing existing state-of-the-art methods. An ablation study confirms the module's efficacy, showing that the ReMA-MobileViT model, integrated with the ReMA module, improved accuracy, recall, and F1-Score by 1.58, 1.76, and 1.58 percentage points, respectively, over the baseline MobileViT. Comparative analyses further illustrate ReMA-MobileViT's superior overall performance; it exceeds classic lightweight MobileNetV2 by 15.77 percentage points and the mainstream Vision Transformer by 2.96 percentage points in accuracy. Critically, ReMA-MobileViT achieves this with significantly fewer model parameters and reduced computational complexity compared to Vision Transformer, establishing a superior balance between accuracy and efficiency. Conclusion The proposed ReMA-MobileViT model offers an effective and lightweight solution for improving sugarcane leaf disease recognition accuracy, particularly in challenging complex backgrounds. Its ability to balance high accuracy with computational efficiency presents a promising technical avenue and a deployable solution for high-precision crop disease diagnosis systems on resource-constrained mobile or edge platforms.

Why it matches plant phenotyping methodsサトウキビ葉の病害状態を画像から認識する軽量深層学習モデルを開発し、精度・計算量・アブレーションを評価しており、植物表現型取得・判定手法が中心である。

abstractWe propose the ReMA-MobileViT model, which significantly enhances feature representation by incorporating a newly designed Residual Multi-head Attention (ReMA) module.

AF-RT-DETR: Adaptive cross-scale feature interaction for real-time plant disease detection in complex field environments →

Introduction Accurate plant disease identification is of great importance for ensuring agricultural productivity and food security. However, complex illumination variations, leaf occlusion, and diverse disease spot scales throughout plant growth stages significantly increase the difficulty of real-time detection, leading to limited accuracy and robustness in existing approaches. Methods To address these challenges, we propose an improved RT-DETRv2-based plant disease detection model, termed AF-RT-DETR. A Bidirectional Cross Gate (BCG) module is introduced in the feature extraction stage to reduce channel redundancy and enhance discriminative feature representation through multi-level feature interactions. The original RepVGG structure is replaced with a Dynamic Channel Shift (DCS) module, effectively enlarging the receptive field and strengthening contextual feature fusion without additional computational overhead. Additionally, an improved Scale-aware Multi-level Loss (SML) emphasizes low-quality feature maps to improve detector robustness. Results The model achieves mAP50 and mAP50:95 of 93.6% and 67.2% on the Plant-Disease dataset, surpassing the baseline by 5.1% and 4.5%. Furthermore, the model was evaluated on multiple crops and growth stages under diverse field conditions, demonstrating robust performance and adaptability. Discussion These results indicate that AF-RT-DETR effectively enables real-time plant disease detection in complex field environments.

Why it matches plant phenotyping methods植物の病徴を画像から検出するモデルの開発と、複数作物・生育段階・圃場条件での性能評価が中心であり、植物病害状態の表現型計測手法に該当する。

abstractwe propose an improved RT-DETRv2-based plant disease detection model, termed AF-RT-DETR.

ODANet: an occlusion and density aware network for small object detection of coffee cherry ripeness in complex field environments. →

Introduction Coffee cherry ripeness assessment is critical for harvesting efficiency and product quality, yet traditional manual inspection methods suffer from subjectivity and low efficiency. Methods To address the challenges of detecting small, occluded, and densely distributed coffee cherries in complex field environments, this study proposes an Occlusion and Density Aware Network (ODANet). Built upon the YOLOv8 framework, ODANet integrates three innovative modules: (1) Condition-Guided Windowed Attention (CGWA), which incorporates occlusion and density maps as auxiliary guidance signals for efficient feature enhancement; (2) Attention-guided Space-Preserving Convolution (ASPC), which employs space-to-depth transformation with cascaded attention to preserve spatial information during downsampling; and (3) Dual-Adaptive Dynamic Upsampling (DADU), which achieves content-adaptive feature reconstruction through dual-branch offset prediction with learnable fusion weights. Results Comprehensive evaluation on a publicly available dataset demonstrates that ODANet achieves state-of-the-art performance among 17 diverse detection architectures, attaining 76.7% mAP@0.5 with a 6.3 percentage point improvement over baseline YOLOv8, while maintaining computational efficiency (8.1 GFLOPs, 30.4M parameters) suitable for real-time deployment. Ablation studies validate the contributions of each module: ASPC improves performance by 2.2%, DADU by 0.6%, and CGWA by 3.5%. Discussion The model demonstrates robust performance across varying lighting conditions, occlusion levels, and growth stages, making it particularly suitable for practical agricultural deployment. This research provides an efficient solution for small object detection in precision agriculture.

Why it matches plant phenotyping methodsコーヒーチェリーの成熟度という植物器官の状態を画像から推定する検出手法を開発し、複数モデル比較・アブレーションで技術的に検証しているため、植物フェノタイピング手法が中心である。

abstractthis study proposes an Occlusion and Density Aware Network (ODANet)

Precise leaf damage detection across diverse species and environments via a large-scale vision model. →

Precise detection of crop leaf damage is essential for real-time plant health monitoring and yield estimation. However, conventional deep learning models often exhibit poor generalization when deployed across varying species and complex, unstructured field environments. To address these limitations, we propose a new modeling paradigm that shifts from traditional task-specific training to foundation model adaptation. Specifically, we introduce a novel architecture integrating the DinoV3 foundation model with a Unet framework to achieve robust leaf lesion segmentation. By incorporating a Spatial Prior Module (SPM) and a Projection Module, our approach effectively bridges the gap between general-purpose pre-training and domain-specific requirements. Experimental results on coffee and black gram datasets demonstrate that this paradigm consistently outperforms standard networks, including Unet, Unet++, and SwinUnet. On the coffee leaf dataset, the proposed model achieves an Intersection over Union (IoU) of 78.31% and a Pixel Accuracy of 88.00%, surpassing the baseline Unet by over 10.5% in IoU. Remarkably, the architecture reduces inference time by approximately 93.6% (from 63.41s to 4.07s), proving that high-parameter foundation models can be adapted for extreme computational efficiency in agricultural scenarios. To further validate scalability, we conduct additional experiments on a larger dataset, AMG HS . The proposed paradigm achieves the best overall detection performance while maintaining superior computational efficiency, confirming its robustness under increased data scale. Interpretability analysis reveals that the foundation model backbone effectively captures high-level semantic features of lesions, providing a clear explanation for its superior performance and cross-domain reliability. This research establishes a scalable, high-performance paradigm for intelligent crop protection, demonstrating that coupling customized encoders with foundation models is a superior strategy for cross-domain agricultural tasks.

Why it matches plant phenotyping methods植物葉の病変を画像からセグメンテーションし、葉の損傷状態を定量化する手法の開発・検証が研究の中心であるため。

abstractwe introduce a novel architecture integrating the DinoV3 foundation model with a Unet framework to achieve robust leaf lesion segmentation.

YOLOv11n-DualPC-Lite: a lightweight, high-precision real-time detection model for maize leaf diseases. →

To address the challenge of balancing model lightweight and detection accuracy in maize leaf disease detection, as well as the limitations of edge device deployment resources, we propose an enhanced target detection model, YOLOv11n-DualPC-Lite.Firstly, the C2fDualPConv module was designed, integrating PartialConv to replace some C3k2 modules in the backbone and neck networks. This approach enhances feature representation while reducing the number of parameters. Secondly, the Slim-Neck architecture is introduced in the neck network. To improve accuracy without increasing the number of parameters, the VoVGSCSPC_SimAm module enables the new Slim-Neck structure to reduce parameters while strengthening feature representation. Finally, an EfficientHead detection head is introduced that uses an inverted bottleneck MBConv module to improve performance. This significantly reduces computational load while efficiently extracting features. This study constructed a maize leaf disease dataset integrating a publicly available Kaggle dataset and a field-collected dataset from Anhui Science and Technology University's experimental plots. The dataset includes four categories: Blight, Common_Rust, Gray_Leaf_Spot, and Health. Through techniques such as rotation and gamma correction, the dataset was expanded from 3,876 to 5,165 images for model training and performance validation. Test results show this improved model performs better than other popular lightweight models overall, with a mAP50 score of 90.9%. Meanwhile, the model has only 2.13 million parameters; its computational complexity is reduced to 4.55 G, and the model size is 4.41 MB. Compared with the original YOLOv11n, its mAP50 is 1.9% higher, while the number of parameters is down by 17.8%, computational complexity is cut by 29.3%, and file size is reduced by 15.7%. When run on a Raspberry Pi 5, the model's detection speed reaches 2.3 FPS, an increase of 27.8%. This model achieves a good balance between detection accuracy and lightweight performance for maize leaf diseases, providing an efficient and practical method for real-time crop disease monitoring.

Why it matches plant phenotyping methodsトウモロコシ葉の病徴を画像から検出・分類する軽量モデルを開発し、データセット構築、性能比較、エッジデバイス検証まで行っており、植物病害状態の画像ベース表現型取得が中心である。

abstractwe propose an enhanced target detection model, YOLOv11n-DualPC-Lite

Adaptive multi-scale feature refinement for wheat phenology recognition using cross-scale attention mechanisms. →

Accurate delineation of crop growth stages under real-world field conditions remains a long-standing challenge in computational phenotyping, particularly for wheat whose developmental phases are characterized by subtle, continuous morphological transitions and environmental noise. In this study, we propose AMFR-Net, an Adaptive Multi-Scale Feature Refinement Network tailored for fine-grained wheat stage identification using ground-level RGB imagery. Unlike conventional architectures that struggle with ambiguous inter-stage boundaries and rigid receptive structures, AMFR-Net leverages a ResNet-101 backbone augmented by a novel Adaptive Multi-Scale Attention Fusion (AMSAF) module-comprising cross-scale interaction blocks and confidence-weighted feature aggregation-to hierarchically recalibrate spatial-semantic representations. This design enables the network to adaptively amplify phenologically salient cues while suppressing irrelevant context, ensuring robust generalization under constrained annotation and deployment conditions. Evaluated on the expert-labeled CGIAR benchmark, AMFR-Net achieves state-of-the-art performance across all major metrics (Top-1 Accuracy: 89.10%; Macro-F1: 89.10%; AUC: 97.88%) and demonstrates superior discriminability in phenologically adjacent stages compared to lightweight and deep CNN baselines. Ablation studies validate the synergistic effect of multi-level attention and scale-aware refinement. The proposed framework offers a scalable, interpretable, and field-deployable solution for in-situ phenology monitoring, and sets a foundation for future integration of multimodal sensing, weak supervision, and cross-seasonal adaptation.

Why it matches plant phenotyping methods小麦の生育ステージを地上RGB画像から推定する新規深層学習手法を開発し、ベンチマーク、比較、アブレーションで検証しており、植物フェノタイピング手法が研究の中心です。

abstractwe propose AMFR-Net, an Adaptive Multi-Scale Feature Refinement Network tailored for fine-grained wheat stage identification using ground-level RGB imagery.

AutoSiQ: a curated haploid Arabidopsis thaliana inflorescence dataset with a fine-grained silique ontology and a deep learning application for haploid fertility quantification. →

Doubled haploid (DH) technology can fast-track crop breeding. Haploid induction yields haploids with only one set of genomes, which are usually sterile. Haploid fertility (HF) is the ability of haploid plants to set seed, and it is a critical bottleneck in DH pipelines. Genetic mechanisms to restore HF hold immense potential in DH crop breeding, yet its phenotyping remains manual, destructive, and inconsistent. While recent advances in imaging and machine learning have improved throughput for general plant traits, no curated image dataset exists for Arabidopsis thaliana that explicitly represents HF. Here, we present AutoSiQ, a dataset and baseline deep learning pipeline for automated HF quantification. AutoSiQ includes high-resolution scanned inflorescences annotated with a seven-class ontology encompassing green siliques, green fertile siliques, mature siliques, fertile siliques, cracked fertile siliques, cracked siliques, and flowers. This multi-class annotation scheme preserves biologically meaningful information beyond binary fertile/non-fertile distinctions, enabling reliable fertility estimation and future phenotyping applications. We release baseline object detection models (YOLOv5), trained using the AutoSiQ dataset, and evaluate their performance across confidence thresholds. Model predictions strongly correlate with manual counts, achieving R² up to 0.94 for total silique number estimation. We further demonstrate AutoSiQ's utility for automated haploid fertility rate (HFR) estimation and genotype discrimination between two contrasting genotypes (WT and bmf2 mutant). A longitudinal analysis identifies ~60 days after sowing (DAS) as the optimal harvest time for maximizing mature silique counts by balancing between the number of immature buds and silique shattering. By releasing both the dataset and baseline code, AutoSiQ provides a reproducible and extensible foundation for high-throughput fertility phenotyping in haploid Arabidopsis .

Why it matches plant phenotyping methodsハプロイド稔性を画像から定量するデータセットと深層学習パイプラインを開発・評価しており、植物フェノタイピング手法が中心である。

abstractHere, we present AutoSiQ, a dataset and baseline deep learning pipeline for automated HF quantification.

CMNet: an asymmetric dual-branch network for accurate cotton segmentation. →

In agricultural automation, precise cotton segmentation is a key step for tasks such as intelligent harvesting and yield estimation. However, in complex field environments, factors such as background interference and irregular target shapes severely affect segmentation accuracy. Existing deep learning methods offer certain advantages but still generally suffer from limitations including insufficient accuracy, over-segmentation, and misidentification. To address these challenges, this study proposes a novel dual-branch cotton segmentation network, Cotton-aware Mamba-enhanced UNet (CMNet), which optimizes the ParaTransCNN architecture by incorporating the 2D Selective Scan (SS2D) module to replace the original Transformer branch, effectively balancing the extraction of local details and global semantic information while reducing computational burden. To enhance the model's perception of irregularly shaped cotton, a Deformable Convolutional Networks v1 (DCNv1) module is integrated into the Vision Mamba (VMamba) branch, further improving the delineation of target boundaries. Additionally, an Atrous Spatial Pyramid Pooling (ASPP) module is introduced at the end of the Convolutional Neural Network (CNN) branch to strengthen multi-scale feature representation. To optimize the fusion of channel and spatial information, the Spatial and Channel Squeeze-and-Excitation (scSE) attention mechanism replaces the original module, enhancing feature modeling capability. Experimental results on an in-field cotton image dataset demonstrate that CMNet outperforms existing mainstream methods, achieving Dice, mIoU, and Accuracy of 91.06%, 84.18%, and 98.10%, respectively, while reducing parameter count and computational complexity, thus exhibiting excellent performance. Furthermore, generalization experiments on multiple other plant datasets also achieved outstanding results, validating the model's adaptability and potential for broader applications in multi-crop segmentation tasks, providing valuable insights for smart agriculture segmentation research. The source code and dataset of this work are publicly available at https://github.com/halidanmu/CMNet.git.

Why it matches plant phenotyping methods綿花画像から植物領域を抽出する新規セグメンテーション手法を中心に開発・検証しており、植物表現型の画像取得・抽出ワークフローに該当する。

abstractthis study proposes a novel dual-branch cotton segmentation network, Cotton-aware Mamba-enhanced UNet (CMNet)

Modeling grain biochemical composition traits of commercial sorghum hybrids under diverse management practices. →

Introduction Sorghum ( Sorghum bicolor (L.) Moench) is a vital cereal crop for food, feed, and biofuel production. Accurate estimation of grain biochemical composition, crude protein (CP), lysine from grain (LysG) and protein (LysP), starch (SC), amylose from grain (AMLG) and starch (AMLS), and crude fat (CF), is crucial for improving breeding and management strategies. Our aim is not pre-harvest forecasting but reducing laboratory cost by identifying a minimal set of post-harvest measurements required to estimate other grain composition traits accurately. Methods We used machine learning (ML) models to predict grain quality traits in commercial sorghum hybrids under different management practices, including precision nitrogen application, cover cropping, and no-till methods. Multi-year field trials (2023-2024) in Saint Charles, Missouri, integrated agronomic, physiological, UAV-based, and environmental data for model training and validation. Results Phenotypic analysis showed that grain composition traits varied significantly by year and management practices. Among ML models, LASSO and ElasticNet achieved the highest predictive accuracy for crude protein (R² = 0.90) and amylose content (AMLS, R² = 0.99; AMLG, R² = 0.92). Bayesian Ridge was most effective for lysine from protein (R² = 0.64), while Partial Least Squares (PLS) excelled in starch content prediction (R² = 0.80). The correlation between grain composition (LysP, CF) and photosystem II efficiency (PhiPS2) indicated that enhanced photosynthesis and yield promote their accumulation. However, Partial Dependence Plots (PDPs) revealed strong non-linear effects, where slight variations in leaf temperature (Tleaf) and stomatal conductance (gsw) were associated with significant shifts in amylose content. Discussion This study highlights the role of genotype × management interactions in sorghum breeding and demonstrates the value of integrating ML-driven models to enhance grain quality and precision agriculture strategies.

Why it matches plant phenotyping methods穀粒の生化学的形質を少数の測定値から推定する機械学習モデルの開発・検証が研究の中心であり、単なる農業実験の routine 測定ではない。

abstractreducing laboratory cost by identifying a minimal set of post-harvest measurements required to estimate other grain composition traits accurately

A digital twin-driven deep learning framework for online quality inspection in tobacco transplanting. →

Tobacco transplanting quality inspection is crucial for tobacco production, as it directly affects crop yield and quality of tobacco leaves. Accurate transplanting status detection and assessment provide essential support for replanting decisions and transplanting machine optimization. Traditional methods rely on manual inspection, which suffer from high cost, low efficiency, and unstable results. To tackle the aforementioned issues, this paper proposes a Deep Learning and Digital Twin driven Online Quality Inspection Method for Tobacco Transplanting, which consists of four core modules: Transplanting Status Detection, Multi-sensor Data Fusion, Digital Twin Visualization, and Operational Optimization Feedback. This paper proposes a lightweight improved YAN-YOLO11 algorithm capable of assessing normal, exposed-root, and buried seedlings. By fusing GNSS positioning data with visual detection results, the system estimates in-row spacing and performs status assessment for missed planting and double planting. The system establishes a virtual-real interactive closed-loop of "collection-detection-mapping-feedback" via the digital twin. By visualizing operational status in real-time and generating replanting path suggestions, it provides guidance for operation management and significantly improves inspection efficiency. Field experiments demonstrate that, compared with YOLO11n, YAN-YOLO11 improves precision and recall by 2.4% and 2.5%, respectively; mAP@50 increased by 3% to 80.9% ± 1.4%, and mAP@0.5:0.95 increased by 5.8% to 54.2% ± 1.0%, while significantly reducing model complexity. The system achieves a real-time performance of 30 FPS in the field, with an overall recognition accuracy of 90.74%, meeting practical application requirements. This study effectively enhances the digitalization, automation, and refined management of tobacco transplanting operations, providing a theoretical foundation and practical solution for the intelligent transformation of transplanting machinery and precision crop management.

Why it matches plant phenotyping methodsタバコ苗の植栽状態、株間、欠株・二株植えを画像とセンサーで推定する手法を開発・検証しており、植物状態の取得が研究の中心である。

abstractThis paper proposes a lightweight improved YAN-YOLO11 algorithm capable of assessing normal, exposed-root, and buried seedlings.

DCSFormer: a high-precision method for cotton seedling point cloud organ segmentation. →

Introduction: Accurately segmenting cotton seedling organs from 3D point clouds is fundamental for high-throughput plant phenotyping and digital breeding. However, cotton seedling segmentation remains challenging due to fine-scale and complex organ morphology, uneven point density with noise, and the lack of high-quality annotated datasets. Methods: To address these issues, we propose DCSFormer, a tailored extension of Point Transformer V3 designed for cotton seedling point cloud segmentation. The model introduces the DCS Block, which leverages dynamic sparse expert routing and dual-channel attention to adaptively capture global semantic dependencies and subtle local geometric variations, thereby improving stem-leaf boundary discrimination. In addition, the proposed CLFSkip replaces traditional skip connections with a cross-layer fusion strategy, effectively integrating multi-scale features while preserving organ-level details. We also constructed an annotated cotton seedling dataset to support training and evaluation. Results and Discussion: Experimental results show that DCSFormer achieves 93.67% mIoU, 95.83% mPrec, 97.35% mRec, and 96.56% mF1, outperforming multiple comparison models. Furthermore, when evaluated against baseline models on two public datasets, Crops3D and Pheno4D, DCSFormer exceeds the baseline across all four metrics, further validating its effectiveness and generalizability. This work provides an effective solution for precise cotton seedling organ segmentation.

Why it matches plant phenotyping methods綿花幼苗の3D点群から器官を抽出する手法を開発し、アノテーション済みデータセットの構築と複数データセットでの性能検証を行っており、植物表現型取得が中心である。

abstractAccurately segmenting cotton seedling organs from 3D point clouds is fundamental for high-throughput plant phenotyping and digital breeding.

Chat Demeter: a multi-agent system for plant disease diagnosis integrating CNN-transformer models →

Plant diseases remain a significant challenge in global agricultural production. Achieving efficient and accurate disease detection is essential for reducing crop losses, controlling agricultural costs, and improving yields. As agriculture rapidly advances toward digitalization and intelligent transformation, the application of artificial intelligence technologies has become a key pathway to enhancing industrial competitiveness. In this study, Chat Demeter, a multi-agent system for plant disease diagnosis based on deep learning. The system captures real-time leaf images through camera devices. It employs a CNN-Transformer model to perform instance segmentation and object detection, thereby enabling automatic identification of diseased leaves and classification of disease types. To enhance interactivity and practical value, the system incorporates a natural language interface, allowing users to upload images and receive automated diagnostic results and treatment suggestions. Experimental results demonstrate that the system achieves an accuracy of 99.50% and an AUC o f 99.91% on the validation dataset, highlighting its superior performance. Overall, Chat Demeter provides an effective tool for crop health monitoring and disease intervention, while offering a feasible pathway and developmental direction for integrating and optimizing future agricultural multi-agent systems.

Why it matches plant phenotyping methods植物葉の画像から病葉をセグメンテーション・分類する診断システムが研究の中心であり、植物病害状態の画像ベース表現型計測に該当する。

abstractChat Demeter, a multi-agent system for plant disease diagnosis based on deep learning.

Agentic AI for smart and sustainable precision agriculture. →

Introduction Ensuring smarter and more sustainable farming practices is a critical challenge in modern agriculture. Agentic Artificial Intelligence (AAI), combined with Precision Agriculture (PA) and Federated Learning (FL), has the potential to enhance decision-making, optimize resource utilization, and reduce environmental impact. Methods This study proposes an AAI based framework for precision agriculture that integrates distributed sensing devices, intelligent agents, and federated learning to enable real time monitoring and decision support at the farm level. A practical deployment architecture is outlined, detailing inter-device communication and localized intelligence. The proposed model is evaluated across two distinct datasets tomato disease classification and weed detection. The model is designed to have DenseNet121, MobileNetV2, EfficientDet-D0, and YOLOv8 as local models within a federated learning environment. Results The federated global model achieved an accuracy of 96.4%, outperforming individual client models, with DenseNet121 and MobileNetV2 attaining accuracies of 95.0% and 93.9%, respectively. For weed species detection, EfficientDet-D0 demonstrated superior performance, achieving an mAP@0.5 of 0.978, average precision of 0.865, and an F1-score of 0.961, compared to YOLOv8 with an mAP@0.5 of 0.956 and an F1-score of 0.935. Discussion The results confirm the feasibility and effectiveness of integrating AAI with federated learning for intelligent precision agriculture. A SWOT analysis highlights the strengths of the proposed approach, along with deployment challenges and constraints. Overall, this study establishes a roadmap for future research, emphasizing sustainable intelligent farming systems.

Why it matches plant phenotyping methods植物病害分類を含む連合学習・エージェント型AI基盤を提案し、データセット上で性能評価しているため、植物の病害状態を推定する計算的フェノタイピング手法が中心である。雑草検出も含まれるが、病害分類の技術評価が明示されている。

abstractThis study proposes an AAI based framework for precision agriculture that integrates distributed sensing devices, intelligent agents, and federated learning to enable real time monitoring and decision support at the farm level.

Lightweight MSW-YOLOv8n-Seg: the instance segmentation of maturity on cherry tomato with improved YOLOv8n-Seg. →

Introduction Automatic and accurate segmentation of cherry tomato maturity in natural environment is the foundation for automatic picking. Lacking of significant differences in adjacent maturity and the problem of mutual occlusion between fruits usually affect the picking process. According to the changes in phenotypic characteristics of cherry tomato during its mature period and the Chinese national standard GH/T 1193-2021, a lightweight maturity instance segmentation method of cherry tomato with 5 levels, including green, turning, pink, light red and red was proposed based on improved YOLOv8n-Seg model, named as MobileViTv3-SK-WIoU-YOLOv8n-Seg (MSW-YOLOv8n-Seg). Methods In this model, MobileViTv3 was introduced into the original YOLOv8 model as backbone for feature extraction to reduce the parameters of the original model; selective kernel (SK) attention module was added to the neck part to improve the feature expression ability of the model; the complete intersection over union (CIoU) loss function in the original head part was replaced with wise intersection over union (WIoU), which can effectively filter low-quality samples and improve the stability and reliability of the model in complex scenes. The proposed model can better balance the relationship between segmentation speed, accuracy, and model computational complexity. Results The experimental results show that the bounding box precision, recall and mean average precision (mAP)@0.5 of the improved model on the test sets were 90.8%, 86.3% and 83.9% respectively, and the model size was 6.0 MB. Compared with YOLOv7-Mask, YOLOv8n-Seg, YOLOv9s-Seg, YOLO11n-Seg, Mask R-CNN (Mask region-based convolutional neural network) and Mask2Former, the bounding box precision increased by 9.6%, 5.2%, 5.7%, 12.3%, 13.3% and 5.0%, the recall increased by 7.8%, 7.4%, 8.8%, 13.1%, 13.9% and 0.1%, and the mAP@0.5 increased by 10.5%, 3.0%, 0.9%, 15.0%, 13.8% and 1.4% respectively. In terms of inference speed, the MSW-YOLOv8n-Seg has the highest inference speed, with FPS of up to 52.9 f·s -1 and latency of only 18.2ms, which demonstrates its real-time processing capability. Discussion The results show that the improved MSW-YOLOv8n-Seg model is optimal, and it suitable for instance segmentation scenarios with high real-time performance and can provide effective exploration for automated cherry tomato fruit picking.

Why it matches plant phenotyping methodsチェリートマト果実の成熟度という植物状態を画像から推定するインスタンスセグメンテーション手法を開発し、精度・速度・モデルサイズを比較検証しており、表現型取得法が中心である。

abstracta lightweight maturity instance segmentation method of cherry tomato with 5 levels, including green, turning, pink, light red and red was proposed based on improved YOLOv8n-Seg model

MTMEGPS: An R package for multi-trait and multi-environment genomic and phenomic selection using deep learning. →

Genomic and phenomic selection have transformed modern breeding by enabling data-driven prediction of complex traits. Deep learning (DL) can further enhance predictive ability by capturing nonlinear patterns that classical and Bayesian approaches often fail to represent. However, despite its potential, the adoption of DL in breeding programs remains limited due to its computational demands and the lack of accessible tools for users without extensive programming experience. This study introduces the MTMEGPS (Multi-Trait and Multi-Environment Genomic and Phenomic Selection), an R package that provides a streamlined end-to-end workflow for Uni- and Multi-Trait (UT and MT, respectively) and Uni- and Multi-Environment (UE and ME, respectively) genomic and phenomic prediction. The package supports data preparation, hyperparameter optimization, model training, and DL-based evaluation. To assess its performance, MTMEGPS was applied to the two default datasets included in the package: Maize (genomic data) and Eucalyptus (near-infrared spectroscopy, NIR, data), as well as to an independent publicly available multi-environment validation dataset. Across most scenarios, MTMEGPS showed superior predictive ability compared with all benchmark models, particularly under UT for the internal datasets and MT for the independent multi-environment dataset. Mean squared error (MSE) values were similar across models, all falling within a moderate range. Overall, these results demonstrate the efficiency and practical utility of MTMEGPS for genomic and phenomic selection, even in scenarios where prediction errors remain moderate.

Why it matches plant phenotyping methods植物の複雑形質を予測するゲノム・フェノミック選抜用Rパッケージを開発し、データ準備からモデル評価までの再利用可能なワークフローを提供・検証しているため、フェノタイピング関連ソフトウェアとして中心的です。

abstractThis study introduces the MTMEGPS (Multi-Trait and Multi-Environment Genomic and Phenomic Selection), an R package that provides a streamlined end-to-end workflow for Uni- and Multi-Trait (UT and MT, respectively) and Uni- and Multi-Environment (UE and ME, respectively) genomic and phenomic prediction.

CNNAttLSTM: an attention-enhanced CNN-LSTM architecture for high-precision jackfruit leaf disease classification. →

Introduction Jackfruit cultivation is highly affected by leaf diseases that reduce yield, fruit quality, and farmer income. Early diagnosis remains challenging due to the limitations of manual inspection and the lack of automated and scalable disease detection systems. Existing deep-learning approaches often suffer from limited generalization and high computational cost, restricting real-time field deployment. Methods This study proposes CNNAttLSTM, a hybrid deep-learning architecture integrating Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) units, and an attention mechanism for multi-class classification of algal leaf spot, black spot, and healthy jackfruit leaves. Each image is divided into ordered 56×56 spatial patches, treated as pseudo-temporal sequences to enable the LSTM to capture contextual dependencies across different leaf regions. Spatial features are extracted via Conv2D, MaxPooling, and GlobalAveragePooling layers; temporal modeling is performed by LSTM units; and an attention mechanism assigns adaptive weights to emphasize disease-relevant regions. Experiments were conducted on a publicly available Kaggle dataset comprising 38,019 images, using predefined training, validation, and testing splits. Results The proposed CNNAttLSTM model achieved 99% classification accuracy, outperforming the baseline CNN (86%) and CNN-LSTM (98%) models. It required only 3.7 million parameters, trained in 45 minutes on an NVIDIA Tesla T4 GPU, and achieved an inference time of 22 milliseconds per image, demonstrating high computational efficiency. The patch-based pseudo-temporal approach improved spatial-temporal feature representation, enabling the model to distinguish subtle differences between visually similar disease classes. Discussion Results show that combining spatial feature extraction with temporal modeling and attention significantly enhances robustness and classification performance in plant disease detection. The lightweight design enables real-time and edge-device deployment, addressing a major limitation of existing deep-learning techniques. The findings highlight the potential of CNNAttLSTM for scalable, efficient, and accurate agricultural disease monitoring and broader precision agriculture applications.

Why it matches plant phenotyping methods植物葉の病害状態を画像から分類する深層学習手法を提案・比較評価しており、病害フェノタイピングの方法開発が中心である。

abstractThis study proposes CNNAttLSTM, a hybrid deep-learning architecture integrating Convolutional Neural Networks (CNN), Long Short-Term Memory (LSTM) units, and an attention mechanism for multi-class classification of algal leaf spot, black spot, and healthy jackfruit leaves.

Multispectral imaging and automated analysis for quantifying grain quality to reveal known and potential novel alleles affecting grain traits in wheat. →

To accelerate the pace of wheat ( Triticum aestivum L.) improvement worldwide, desired seed-level characteristics and seed quality receive a growing attention as they directly impact early seedling establishment, seed longevity, and grain quality. Nevertheless, the throughput and accuracy of seed-level phenotyping and analysis have become a key limiting factor in this research domain, requiring new solutions to relieve this bottleneck. In this study, we first combined automated multispectral seed imaging (MSI; i.e. the VideometerLab 4 and Autofeeder systems) with a variety of machine learning and computer vision techniques to establish a high-throughput pipeline to analyse wheat seeds. Then, using 493 lines selected from the NIAB Diverse MAGIC (NDM) population, we applied the pipeline to segment individual seeds from MSI seed-lot images. This enabled us to perform seed-level measurement of sixteen morphological (e.g. seed size, length, width, and roundness) and spectral traits, ranging from ultraviolet (i.e. 375 nm, correlating with crude protein) to near-infrared (e.g. 975 nm, for assessing water content) wavelengths. After verifying these seed quality related traits (R2 ≥ 0.949; p < 0.001), we applied genome-wide association studies (GWAS) to link the computationally derived traits to genetic loci and identified eleven significant loci. Some of the loci were previously reported, with two unknown loci valuable for further assessment. Taken together, we believe this integrated MSI analysis pipeline provides a powerful solution for seed research and crop improvement in wheat, enabling us to bridge MSI, seed-level analysis, and genetic mapping to assess seed morphology, seed quality, and their underlying genetic architectures effectively.

Why it matches plant phenotyping methods自動マルチスペクトル画像と機械学習・コンピュータビジョンを統合し、個々の小麦種子の形態・スペクトル形質を高スループットに抽出するパイプラインが研究の中心である。

abstractwe first combined automated multispectral seed imaging (MSI; i.e. the VideometerLab 4 and Autofeeder systems) with a variety of machine learning and computer vision techniques to establish a high-throughput pipeline to analyse wheat seeds.

Electrochromic polyoxometalates for sensing abiotic stress in plants. →

Introduction Understanding plant responses to abiotic stress requires an insight into plant redox activity. This study proposes a novel and cost-effective method for assessing the redox state of plants. Methods The method utilizes the electrochromic properties of polyoxometalate phosphomolybdic acid hydrate (PMA). PMA is reduced proportionally by glutathione (GSH) and ascorbic acid (AsA), which results in a measurable color change. The validity of this method was confirmed through empirical experimentation in Arabidopsis thaliana under conditions of salinity and UV radiation. Results Salinity treatments revealed a non-significant, two-phase trend in redox activity with an increase at moderate levels followed by a decrease. UVC radiation led to a substantial decrease in redox activity, indicating distress. In contrast, UVA promoted resilience, also known as eustress. Notably, UVB significantly increased redox activity, suggesting the activation of an emergency antioxidant response. Discussion A demonstrable correlation has been identified between the redox activity of plants and various stress types. This correlation facilitates the classification of responses into two distinct categories: adaptive eustress and detrimental distress. This advancement contributes to the enhancement of plant metabolic and stress tolerance evaluation.

Why it matches plant phenotyping methods植物のレドックス状態を測定する新規手法を開発し、シロイヌナズナでストレス条件下の妥当性を検証しており、表現型取得が研究の中心である。

abstractThis study proposes a novel and cost-effective method for assessing the redox state of plants.

Gradient-guided boundary-aware selective scanning with multi-scale context aggregation for plant lesion segmentation →

Introduction Plant lesion segmentation aims to delineate disease regions at the pixel level to support early diagnosis, severity assessment, and targeted intervention in precision agriculture. However, the task remains challenging due to large variations in lesion scale—ranging from minute incipient spots to coalesced regions—and ambiguous, low-contrast boundaries that blend into healthy tissue. Methods We present GARDEN, a Gradient-guided boundary-Aware Region-Driven Edge-refiNement network that unifies multi-scale context modeling with selective long-range boundary refinement. Our approach integrates a Multi-Scale Context Aggregation (MSCA) module to harvest contextual cues across diverse receptive fields, forming scale-consistent lesion priors to improve sensitivity to tiny lesions. Additionally, we introduce a Boundary-aware Selective Scanning (BASS) module conditioned on a Gradient-Guided Boundary Predictor (GGBP). This module produces an explicit boundary prior to steer a Mamba-based 2D selective scan, allocating long-range reasoning to boundary-uncertain pixels while relying on local evidence in confident interiors. Results Validated across two public plant disease datasets, GARDEN achieves state-of-the-art results on both overlap and boundary metrics. Specifically, the model demonstrates pronounced gains on small lesions and boundary-ambiguous cases. Qualitative results further show sharper contours and reduced spurious responses to illumination and viewpoint changes compared to existing methods. Discussion By coupling scale robustness with boundary precision in a single architecture, GARDEN delivers accurate and reliable plant lesion segmentation. This method effectively addresses key challenges in the field, offering a robust solution for automated disease analysis under challenging real-world conditions.

Why it matches plant phenotyping methods植物病斑を画像から分割し、病害状態・重症度を推定する新規手法を開発し、公開データセットで検証しているため、植物フェノタイピング手法が中心である。

abstractPlant lesion segmentation aims to delineate disease regions at the pixel level to support early diagnosis, severity assessment, and targeted intervention in precision agriculture.

Adaptive preprocessing and Cascaded Canny Edge Segmentation for cassava disease identification using HyperCapsInception-ResNet-V2-CNN. →

Introduction Cassava is one of the most widely cultivated crops worldwide, renowned for its rich natural ingredients and numerous nutritional benefits. However, the complex interdependencies among its features often pose challenges in image restoration and segmentation, particularly when identifying disease regions. In previous work, this manifested as higher false positives and misidentification of non-relevant areas, leading to a decline in precision and accuracy. Methods To address these issues, this study proposed an efficient artificial intelligence-powered image analysis system that leverages optimal feature selection with a HyperCapsInception-ResNet-V2-CNN model to enhance disease detection accuracy. Initially, the dataset was collected from the Kaggle repository, its name was Cassava Leaf Disease Classification, and it comprised 21,367 different images. Our approach began by normalizing cassava plant disease data using adaptive Gaussian Otsu thresholding. Histogram color evaluation and iterative clustering fragmentation were then applied to better isolate disease variations and improve precision. Subsequently, Cascaded Canny Edge Segmentation (CCES) was used to effectively segment the disease region. The disease variation properties were further evaluated using the Optimal Spider Swarm Intelligence Technique (OSSIT) to reduce irrelevant feature dimensions. For classification, the HyperCapsInception-ResNet-V2-CNN model was employed to categorize cassava diseases, including cassava bacterial blight (CBB), cassava mosaic disease (CMD), cassava green mite (CGM) disease, and cassava brown streak disease (CBSD), along with regular and abnormal leaf states. Results The proposed method's simulation results achieved 98.15% accuracy, a 97.22% F1-score, and 96.02% precision, outperforming other traditional methods such as EfficientNetB3, AlexNet, Faster-RCNN, and InceptionV3. Discussion Both optimized feature selection with OSSIT and hybrid HyperCapsInception-ResNet-V2-CNN architecture significantly enhanced the detection reluctance and the classification of the data. These findings indicate that the proposed system is effective in the automated detection of cassava disease and has a high potential of being practical in agricultural practices especially in precision farming and early detection of diseases.

Why it matches plant phenotyping methodsカッサバ葉画像から病変領域を分割・抽出し、病害状態を分類する画像解析手法の開発が研究の中心であるため、植物表現型手法として採用。

abstractCascaded Canny Edge Segmentation (CCES) was used to effectively segment the disease region.

Advanced hyperspectral image processing and machine learning approaches for early detection of wheat stem rust. →

Hyperspectral remote sensing has shown great promise for early detection of plant diseases, yet its adoption is often hindered by spectral variability, noise, and distribution shifts across acquisition conditions. In this study, we present a systematic preprocessing pipeline tailored for hyperspectral data in plant disease detection, combining pixel-wise correction, curve-wise normalization and smoothing, and channel-wise standardization. The pipeline was evaluated on an experiment on early detection of stem rust ( Puccinia graminis f. sp. tritici Eriks. and E. Henn.) of wheat ( Triticum aestivum L.). The pipeline implementation enhanced the classification models accuracy raising F1-scores of logistic regression, support vector machines and Light Gradient Boosting Machine from 0.67-0.75 (raw spectra) to 0.86-0.94. Notably, it enabled reliable detection of asymptomatic infections as early as 4 days after inoculation, which was not achievable without preprocessing. The framework demonstrates potential for generalization beyond plant pathology, suggesting applicability to a range of hyperspectral remote sensing tasks such as vegetative health monitoring, environmental assessment, and material classification through improved signal interpretability and robustness. This work lays the groundwork for advancing hyperspectral image processing by proposing a reproducible, scalable pipeline that could be adapted for integration into unmanned and satellite imaging systems.

Why it matches plant phenotyping methods小麦茎锈病の無症状感染を対象に、ハイパースペクトル画像の前処理パイプラインを開発・評価し、植物病害状態の早期推定性能を検証しているため、植物フェノタイピング手法が中心である。

abstractwe present a systematic preprocessing pipeline tailored for hyperspectral data in plant disease detection

Comprehensive AI framework for automated classification, detection, segmentation, and severity estimation of date palm diseases using vision-language models and generative AI. →

The date palm ( Phoenix dactylifera L.) is a vital crop in arid and semi-arid regions, contributing over $13 billion annually to the global economy. However, it faces significant yield losses due to pests, such as the red palm weevil, and diseases, including Bayoud and Black Scorch. Currently, expert visual inspection is the primary method of management, but it is time-consuming, subjective, and unsuitable for detecting large-scale or early-stage damage. Automated approaches based on classical machine learning offer limited improvements due to their lack of generalizability and environmental sensitivity. Recent deep learning methods, such as CNNs and Vision Transformers, have improved classification accuracy, but treat tasks like classification, detection, segmentation, and severity estimation as separate. This paper proposes an integrated Reveal-Aware Hybrid Vision-Language and Transformer-based AI framework that combines GAN-based augmentations for feature generation, CLIP for multimodal classification, PaliGemma2 for text-based detection, Grounding DINO + SAM 2.1 for zero-shot segmentation, and a Vision Transformer regression model for severity prediction. This end-to-end explainable diagnostic pipeline achieved 98% classification accuracy, 95.8% precision, 91.3% recall, and 94.2% F1-score across two datasets: nine classes of infected date palm leaves and three classes of date palm diseases. The proposed framework demonstrated detection accuracy of 94-98%, high-quality segmentations, and reliable severity estimates. This integrated approach highlights the potential of combining AI, vision-language models, and transformers for scalable, accurate, and sustainable plant disease management.

Why it matches plant phenotyping methods植物病害の分類・検出・セグメンテーション・重症度推定を統合した画像ベース手法の開発と評価が中心であり、感染植物の状態を直接推定するため収録対象。

abstractThis paper proposes an integrated Reveal-Aware Hybrid Vision-Language and Transformer-based AI framework

Machine learning-enabled UAV hyperspectral identification of tomato spotted wilt virus in tobacco. →

Problems Tomato Spotted Wilt Virus (TSWV) severely affects tobacco yield and quality, creating an urgent need for accurate, rapid, non-destructive monitoring to support disease management. While existing TSWV detection methods perform well at the leaf scale, their field-scale application remains challenging. Due to complex crop canopy structures, spectral characteristics at the field level differ significantly from leaf-level observations, and TSWV-sensitive spectral features are still unclear. This study therefore aims to develop a field-scale TSWV identification model using UAV-based hyperspectral imaging to enable targeted disease control. Methodology A UAV-mounted hyperspectral camera (400-1000 nm) was deployed to capture imagery of tobacco plants at the rosette stage, enabling comparative spectral analysis between healthy and infected specimens. To identify sensitive features associated with tobacco plants infected with TSWV, six distinct feature extraction methodologies encompassing traditional statistical approaches (spectral ratio, correlation analysis, and principal component analysis [PCA]), machine learning-based techniques (relevant features [Relief], successive projections algorithm) and vegetation indices were utilized. Subsequently, we conducted a systematic evaluation of 18 classification models developed using three machine learning algorithms-support vector machine (SVM), k-nearest neighbors, and extreme gradient boosting -with the derived feature variables. Results This study demonstrates that while all integrated models combining Relief- and Correlation- selected feature bands with three machine learning algorithms delivered excellent performance, the SVM-Relief model achieved the most outstanding results (OA = 97.3%, AUC = 0.994, Kappa=0.947). Based on the SVM-Relief combination, a proposed method called RPR -which integrates PCA with recursive feature elimination- was further employed to reduce the number of feature indicators from 15 to 4 (775.6/772.9/781.1/756.4 nm). The resulting SVM-RPR combination model achieved performance (OA = 97.3%, AUC = 0.990, Kappa=0.947) comparable to that of the SVM-Relief model. Contribution This indicated that red-edge bands were of significant value in distinguishing healthy and TSWV-infected tobacco plants. Our study indicates the significant potential of integrating UAV-based hyperspectral imaging with machine learning techniques for rapid, non-destructive detection of tobacco TSWV at the field scale. The proposed approach offers a novel and efficient pathway for remote sensing-based monitoring of viral diseases in crops, with implications for precision agriculture and plant disease management.

Why it matches plant phenotyping methodsUAVハイパースペクトル画像と機械学習により、タバコ個体のウイルス感染状態を圃場規模で推定する手法の開発・評価が研究の中心であるため。

abstractThis study therefore aims to develop a field-scale TSWV identification model using UAV-based hyperspectral imaging to enable targeted disease control.

Real-time segmentation and phenotypic analysis of rice seeds using YOLOv11-LA and RiceLCNN. →

Introduction The real-time, accurate detection and classification of rice seeds are crucial for improving agricultural productivity, ensuring grain quality, and promoting smart agriculture. Although significant progress has been made using deep learning, particularly convolutional neural networks (CNNs) and attention-based models, earlier methods such as threshold segmentation and single-grain classification faced challenges related to computational efficiency and latency, especially in high-density seed agglutination scenarios. This study addresses these limitations by proposing an integrated intelligent analysis model that combines object detection, real-time tracking, precise classification, and high-accuracy phenotypic measurement. Methods The proposed model utilizes the lightweight YOLOv11-LA for real-time grain segmentation, which builds upon the YOLOv11 architecture. YOLOv11-LA incorporates several enhancements over YOLOv11, including separable convolutions, CBAM (Convolutional Block Attention Module) attention mechanisms, and module pruning strategies. These modifications not only improve detection accuracy but also significantly reduce the number of parameters by 63.2% and decrease computational complexity by 51.6%. For classification, the model employs a custom-designed, lightweight RiceLCNN classifier. Additionally, the DeepSORT algorithm is employed for real-time multi-object tracking, and sub-pixel edge detection along with dynamic scale calibration mechanisms are applied for precise phenotypic feature measurement. Results Compared to YOLOv11, the YOLOv11-LA model increases the mAP@0.5:0.95 score by 1.9%, showcasing its superior detection performance while maintaining lower computational overhead. The RiceLCNN classifier achieved classification accuracies of 89.78% on private datasets and 96.32% on public benchmark datasets. The system demonstrated high accuracy in measuring phenotypic features such as seed size and roundness, with measurement errors kept within 0.1 millimeters. The DeepSORT algorithm effectively managed multi-object tracking, reducing duplicate identifications and frame loss in real-time. Discussion Experimental validation confirmed that the YOLOv11-LA model outperforms the original YOLOv11 in terms of both detection speed and accuracy, while also maintaining low computational complexity. The integration of the YOLOv11-LA, RiceLCNN, and DeepSORT algorithms, combined with advanced measurement techniques, underscores the model's potential for industrial applications, particularly in enhancing smart agricultural practices.

Why it matches plant phenotyping methodsイネ種子画像からサイズや真円度を抽出するリアルタイム画像解析手法を開発し、精度・速度・測定誤差を検証しており、表現型取得が研究の中心です。

abstractThe proposed model utilizes the lightweight YOLOv11-LA for real-time grain segmentation

Detection techniques for tomato diseases under non-stationary climatic conditions. →

Tomato growth is highly susceptible to diseases, making accurate identification crucial for timely intervention. While deep learning models like the YOLO family have demonstrated success in detecting diseases in agricultural settings, they typically assume that training and testing data are independently and identically distributed (i.i.d.), which often doesn't hold in real-world scenarios. When pre-trained models are applied to new environments, performance can degrade due to domain shifts. To address this, we propose CTTA-DisDet, a continuous test-time domain adaptation framework for tomato disease detection that adapts models to evolving environments during testing, improving generalization in unseen domains. CTTA-DisDet utilizes a teacher-student architecture where both models share the same structure. Dynamic data augmentation is introduced, involving explicit and implicit augmentations. Explicit augmentation corrupts input images, while implicit augmentation uses large language models (LLMs) to generate new domain data. The teacher model learns generalized knowledge, and the student model mimics the teacher to distill domain-specific information. During testing, pseudo-labels generated by the teacher update the student model. To prevent catastrophic forgetting, a subset of neurons is randomly restored to their original weights during each test-time iteration. The teacher model is continuously updated via exponential moving average (EMA). Experimental results demonstrate that CTTA-DisDet achieves an impressive 67.9% performance in continuously changing cross-domain environments, significantly benefiting practical applications in non-stationary settings.

Why it matches plant phenotyping methodsトマトの病徴を画像から検出するためのドメイン適応手法を提案し、異なる環境で性能評価しており、植物病害状態の表現型取得が中心である。

abstractwe propose CTTA-DisDet, a continuous test-time domain adaptation framework for tomato disease detection that adapts models to evolving environments during testing, improving generalization in unseen domains.

LBS-YOLO: a lightweight model for strawberry ripeness detection. →

Introduction The traditional strawberry picking operation has long relied on manual work. With the aging trend of the population becoming more and more obvious, the application of intelligent picking technology has become an irreversible trend. However, existing recognition methods still face bottlenecks such as suboptimal recognition accuracy and low computational efficiency. To address these issues, this study constructs a lightweight detection model, LBS-YOLO, based on an improved YOLOv11n architecture, significantly the model's accuracy and interference robustness while greatly compressing the parameter quantity. Methods The LBS-YOLO model is built upon YOLOv11n as the baseline network. In order to enhance the ability of backbone network feature representation, the model designs a lightweight LAWDS module. This design combines channel attention with spatial reconstruction operation to optimize the information retention efficiency in the down-sampling process, thus effectively enhancing the multi-scale feature representation ability and gradient flow propagation performance. Then in the feature fusion stage, the model introduces a Bidirectional Feature Pyramid Network (BiFPN), which not only enables cross-scale feature fusion but also achieves adaptive weighting through a learnable weight allocation mechanism. At last, adopts the C3k2_Star module to replace the conventional C3K2 for improved feature representation. Results On the used strawberry dataset, the LBS-YOLO model reached 88.6% mAP@0.5 and 75.8% mAP@0.5:0.95, which were 2.2 and 1.3 percentage points higher than YOLOv11n, respectively. The LBS-YOLO model improves the recall rate from 83.2% of YOLOv11n to 86.4%, and the F1-score from 81.2% to 82.9%. Its computational complexity is 6.6 GFLOPs and its reasoning speed is 260.7 FPS. Even better, LBS-YOLO only needs 3.4MB of storage space and 1.6 million parameters, which are 34.6% and 38% less than YOLOv11n respectively. Discussion The experiment demonstrates that, the LBS-YOLO model can significantly reduce the number of parameters and effectively improve the detection accuracy and operation efficiency. It successfully alleviated the problems of false detection and missed detection, thereby providing reliable technical support for strawberry growth monitoring, maturity identification and automatic picking.

Why it matches plant phenotyping methodsイチゴの成熟度という植物状態を画像から推定する軽量検出モデルを開発・評価しており、フェノタイピング手法が中心である。

abstractthis study constructs a lightweight detection model, LBS-YOLO, based on an improved YOLOv11n architecture

Deep learning-based phenotyping of lettuce diseases using Efficient-FBM-FRMNet for precision agriculture. →

Lettuce ( Lactuca sativa ), a widely cultivated leafy vegetable, is highly susceptible to bacterial and fungal infections that severely reduce yield and quality. Rapid and accurate disease identification is therefore essential for precision agriculture and sustainable crop management. This study proposes Efficient-FBM-FRMNet, a modular deep learning framework for automated lettuce disease detection. The model integrates EfficientNetB4 with dilated convolutions, a Feature Bottleneck Module (FBM) for redundancy reduction, a Reasoning Engine for higher-order semantic inference, and a Feature Refinement Module (FRM) for enhanced generalization. The framework was trained and validated on a publicly available dataset of 2,813 lettuce leaf images (bacterial, fungal, and healthy classes) using stratified 5-fold cross-validation. The proposed Efficient-FBM-FRMNet achieved an overall accuracy of 97.5%, outperforming baseline CNNs such as EfficientNetB4, ResNet50, and DenseNet121. It demonstrated superior precision (96.0%), recall (96.6%), and F1-score (97.0%), confirming its robustness and consistency across multiple folds. Statistical significance analysis (p

Why it matches plant phenotyping methodsレタス葉画像から病害状態を推定する深層学習フレームワークを開発し、公開データセット上で交差検証・ベースライン比較により性能を評価しており、植物表現型取得手法が中心である。

abstractThis study proposes Efficient-FBM-FRMNet, a modular deep learning framework for automated lettuce disease detection.

CottonNet-MHA: a multi-head attention-based deep learning framework for cotton disease detection. →

India is an agro-based country. The major goal of agriculture is to produce disease-free healthy crops. For Indian agronomists, cotton is a profitable commercial and fibre crop, it is the world's second-biggest export crop after China. Cotton production is also affected in a negative way by high use of water, authority of soil erosion and the practice of using dangerous fertilizers and pesticides. The two greatest threats to the rapid growth of the crop are the sucking bugs and cotton diseases. Prompt detection and accurate identification of diseases is vital to ensure healthy crop growth and achieve better yields. The primary objective of this research is to build a model by implementing deep learning-based approaches to spot infections in cotton crops. Deep learning is used because of its exceptional results in classification and image processing tasks. To address this issue, we developed CottonNet-MHA a novel deep learning framework to identify pathological symptoms in cotton leaves. The model employs multi-head attention mechanisms to strengthen feature learning and highlight the diseased-affected regions. To evaluate the performance of the proposed model, five pretrained transfer learning architectures-VGG16, VGG19, InceptionV3, Xception, and MobileNet were used as benchmark models. Furthermore, Gradient-weighted Class Activation Mapping (Grad-CAM) visualization was applied to enhance the trustworthiness and interpretability of the model. A web-based application was developed to deploy the trained model for real-world applicability. The performance analysis is carried out on the developed model based on the conventional models and the results indicate that CottonNet-MHA dominates the conventional models with respect to its accuracy as well as efficiency in the detection of diseases. The use of attention mechanisms approach strengthens the model's diagnostic accuracy and overall reliability. Grad-CAM results further demonstrated that the model effectively targets diseased areas, enhancing interpretability and reliability. Discussion: The study shows that CottonNet-MHA not only automates disease detection but also enhances interpretability through Grad-CAM analysis. The developed web platform allows the model to be applied in real-world environments, supporting live disease monitoring. The proposed framework not only improves the accuracy of cotton disease diagnosis but also offers potential for extension to other crop disease detection systems.

Why it matches plant phenotyping methods綿花葉の病徴を画像から検出・分類する深層学習手法を開発し、既存モデルとの比較検証とGrad-CAMによる病変領域の解釈を行っており、植物病害状態の表現型取得が中心である。

abstractThe primary objective of this research is to build a model by implementing deep learning-based approaches to spot infections in cotton crops.

GAE-YOLO: a lightweight multimodal detection framework for tomato smart agriculture with edge computing →

Introduction The advancement of smart agriculture has witnessed increasing applications of computer vision in crop monitoring and management. However, existing approaches remain challenged by high computational complexity, limited real-time capability, and poor multi-task coordination in tomato cultivation scenarios. Methods To address these limitations, an intelligent tomato management system is proposed based on the Ghost-based Adaptive Efficient You Only Look Once (GAE-YOLO) algorithm. The lightweight architecture of the GAE-YOLO framework is achieved through the replacement of standard convolutional layers with Ghost Convolution (GhostConv) modules, while detection accuracy is significantly improved by the integration of both AReLU activation functions and Effective Intersection over Union (E-IoU) loss optimization. The system, implemented on a Jetson TX2 embedded platform, also incorporates ZED stereo vision for 3D localization and a PyQt6-based visualization platform. Results When implemented on Jetson TX2, the system achieving 93.5% mean Average Precision at 50% intersection over union (mAP@50) at 10.2 frames per second (FPS), which can be optimized to 27 FPS by employing TensorRT acceleration and 720p resolution for scenarios demanding higher throughput. Furthermore, it establishes standardized assessment systems for tomato maturity and yield prediction, and offers integrated modules for disease diagnosis and agricultural large language model consultation. Discussion This work establishes a new paradigm for edge computing in agriculture while providing critical technical support for smart farming development.

Why it matches plant phenotyping methodsトマトの成熟度・収量予測および病害診断を含む画像・3Dビジョン基盤を開発し、エッジ環境で性能評価しているため、植物表現型取得が中心的な研究である。

abstractan intelligent tomato management system is proposed based on the Ghost-based Adaptive Efficient You Only Look Once (GAE-YOLO) algorithm

TflosYOLO+TFSC: an accurate and robust model for estimating flower count and flowering period. →

Tea flowers play a crucial role in taxonomic research and hybrid breeding of tea plants. As traditional methods of observing tea flower traits are labor-intensive and inaccurate, TflosYOLO and Tea Flowering Stage Classification (TFSC) models were proposed for tea flowering quantification, which enable the estimation of flower count and flowering period. In this study, a highly representative and diverse dataset was constructed by collecting flower images from 29 tea accessions in 2 years. Based on this dataset, the TflosYOLO model was built on the YOLOv5 architecture and enhanced with the Squeeze-and-Excitation (SE) network, Adaptive Rectangular Convolution, and Attention Free Transformer, which is the first model to offer a viable solution for detecting and counting tea flowers. The TflosYOLO model achieved a mean Average Precision at 50% IoU (mAP50) of 0.844, outperforming YOLOv5, YOLOv7, and YOLOv8. Furthermore, the TflosYOLO model was tested on 31 datasets encompassing 26 tea accessions and five flowering stages, demonstrating high generalization and robustness. The correlation coefficient (R 2 ) between the predicted and actual flower counts was 0.964. Additionally, the TFSC model-a seven-layer neural network-was designed for the automatic classification of the flowering period. The TFSC model was evaluated for 2 years and achieved an accuracy of 0.738 and 0.899. Using the TflosYOLO+TFSC model, the tea flowering dynamics were monitored, and the changes in flowering stages were tracked across various tea accessions. The framework provides crucial support for tea plant breeding programs and the phenotypic analysis of germplasm resources.

Why it matches plant phenotyping methods茶花画像から花数と開花期を推定するモデルを開発・検証しており、植物表現型の取得・抽出手法が研究の中心である。

abstractTflosYOLO and Tea Flowering Stage Classification (TFSC) models were proposed for tea flowering quantification, which enable the estimation of flower count and flowering period.

VitiForge: a new procedural pipeline approach for grapevine disease identification under data scarcity. →

Early identification of grapevine diseases is critical for reducing yield losses and ensuring sustainable viticulture. CNNs trained on benchmark datasets such as PlantVillage often achieve near-perfect accuracy, yet this performance fails to translate to real-world field conditions where lighting, backgrounds, and lesion appearance vary widely. To address challenges of data scarcity and imbalance, this study introduces VitiForge, a novel procedural synthetic imagery pipeline for generating realistic synthetic grape leaf textures representing healthy, Black Rot, Esca, and Leaf Blight conditions. VitiForge is systematically evaluated against GAN-based augmentation through a data ablation study on PlantVillage and FieldVitis, a curated field dataset, using MobileNetV2, InceptionV3, and ResNet50V2 classifiers. Results show that VitiForge significantly improves performance in low-data regimes, enabling model training even without real samples, whereas GAN augmentation proves more effective once sufficient real data is available. On field imagery, VitiForge often matched or surpassed GAN-based methods, particularly when paired with MobileNetV2. These findings highlight the complementary roles of procedural and GAN-based synthetic data: VitiForge offers flexibility and scalability under cross-domain and data-scarce conditions, while GANs enhance realism and variability when ample data exists. Together, they support the development of robust and generalizable models for automated grape disease detection in precision agriculture.

Why it matches plant phenotyping methodsブドウ葉の病徴を画像から識別するための合成画像生成パイプラインを開発し、GAN augmentationと比較評価しているため、植物病害状態のフェノタイピング手法が中心である。

abstractthis study introduces VitiForge, a novel procedural synthetic imagery pipeline for generating realistic synthetic grape leaf textures representing healthy, Black Rot, Esca, and Leaf Blight conditions.

A hybrid vision transformer and ResNet18 based model for biotic rice leaf disease detection. →

Introduction Agriculture is crucial to human survival. The growing of biotic rice plants is very helpful for feeding a lot of people around the world, especially in places where rice is a main food. The detection of rice leaf disease is critical to increasing crop productivity. Methods To improve the accuracy of rice leaf disease prediction, this paper proposes a hybrid Vision Transformer (ViT) with pre-trained ResNet18 models (ViT-ResNet18). In general, the input images apply to the pre-trained ViT and ResNet18 models independently. The output features of these two models are combined and fed into the final Fully Connected (FC) layer, followed by a Softmax layer for final classification. Results The output of rice leaf diseases from the FC layer of the proposed hybrid ViT with ResNet18 model achieved 94.4% accuracy, a precision of 0.948, a recall of 0.944, an F1-Score of 0.942, and an Area Under Curve (AUC) of 0.985. Discussion The proposed hybrid model ViT-ResNet18 shows a 5%, 1%, and 1% improvement in accuracy compared to VGG16 with Neural Network, Inception V3 with Neural Network, and SqueezeNet with Neural Network classifier, respectively.

Why it matches plant phenotyping methodsイネ葉の病害状態を画像から分類する深層学習手法の提案・比較評価が研究の中心であり、植物病害フェノタイピング手法に該当する。

abstractthis paper proposes a hybrid Vision Transformer (ViT) with pre-trained ResNet18 models (ViT-ResNet18).

DP-MaizeTrack: a software for tracking the number of maize plants and leaves information from UAV image. →

In modern agricultural production, accurate monitoring of maize growth and leaf counting is crucial for precision management and crop breeding optimization. Current UAV-based methods for detecting maize seedlings and leaves often face challenges in achieving high accuracy due to issues such as low spatial-resolution, complex field environments, variations in plant scale and orientation. To address these challenges, this study develops an integrated detection and visualization software, DP-MaizeTrack, which incorporates the DP-YOLOv8 model based on YOLOv8. The DP-YOLOv8 model integrates three key improvements. The Multi-Scale Feature Enhancement (MSFE) module improves detection accuracy across different scales. The Optimized Spatial Pyramid Pooling-Fast (OSPPF) module enhances feature extraction in diverse field conditions. Experimental results in single-plant detection show that the DP-YOLOv8 model outperforms the baseline YOLOv8 with improvements of 3.9% in Precision (95.1%), 4.1% in Recall (91.5%), and 4.0% in mAP50 (94.9%). The software also demonstrates good accuracy in the visualization results for single-plant and leaf detection tasks. Furthermore, DP-MaizeTrack not only automates the detection process but also integrates agricultural analysis tools, including region segmentation and data statistics, to support precision agricultural management and leaf-age analysis. The source code and models are available at https://github.com/clhclhc/project.

Why it matches plant phenotyping methodsUAV画像からトウモロコシ個体数と葉数を抽出するソフトウェアを開発しており、植物形質取得が研究の中心です。

abstractthis study develops an integrated detection and visualization software, DP-MaizeTrack

Research on the intelligent detection model of plant diseases based on MamSwinNet. →

Plant diseases pose a severe threat to global agricultural production, significantly challenging crop yield, quality, and food security. Therefore, accurate and efficient disease detection is crucial. Current detection methods have clear limitations: CNN-based methods struggle to model long-range dependencies effectively and have weak generalization abilities. Transformer-based methods, while adept at long-range feature modeling, face issues with large parameter sizes and inefficient calculations due to the quadratic complexity of the self-attention mechanism in relation to image size. To address these challenges, this paper proposes the MamSwinNet model. Its core innovation lies in: using the Efficient Token Refinement module with an overlapping space reduction method, relying on depthwise separable convolutions designed with “stride + 3” convolution kernels to expand the image block overlap area and fully preserve boundary spatial structure. This generates high-quality tokens and converts them into a fixed number of latent tokens, reducing computational complexity while maximizing the retention of key features. It integrates the Spatial Global Selective Perception (SGSP) module and the Channel Coordinate Global Optimal Scanning (CCGOS) module. The SGSP module uses a dual-branch structure (the spatial modeling branch introduces 2D-SSM to scan four directions for capturing long-range dependencies, and the residual compensation branch supplements features to prevent loss; the two branches are combined using Hadamard product to enhance spatial detail modeling). The CCGOS module combines channel and spatial attention by embedding positional information through global average pooling in the height and width dimensions, using the Mamba block for channel-selective scanning and generating an attention map, enabling precise association of key channel features like color with spatial distribution. Experimental results show that the model achieves F1 scores of 79.47%, 99.52%, and 99.38% on the PlantDoc, PlantVillage, and Cotton datasets, respectively. The model has only 12.97M parameters (52.9% less than the Swin-T model) and a computational cost as low as 2.71GMac, significantly improving computational efficiency. This study provides an efficient and reliable intelligent solution for large-scale crop disease detection.

Why it matches plant phenotyping methods植物病害を画像から検出・判定するMamSwinNetモデルの開発とデータセット評価が研究の中心であり、感染植物の病態を直接推定する画像ベース表現型解析に該当する。

abstractExperimental results show that the model achieves F1 scores of 79.47%, 99.52%, and 99.38% on the PlantDoc, PlantVillage, and Cotton datasets, respectively.

Cross-scale detection and cross-crop generalization verification of tomato diseases in complex agricultural environments. →

In order to overcome the key challenges associated with detecting tomato leaf disease in complex agricultural environments, such as leaf occlusion, variation in lesion size and light interference, this study presents a lightweight detection model called ToMASD. This model integrates multi-scale feature decoupling and an adaptive alignment mechanism. The model innovatively comprises a dual-branch adaptive alignment module (TAAM) that achieves cross-scale lesion semantic alignment via a dynamic feature pyramid, a local context-aware gated unit (Faster-GLUDet) that uses a spatial attention mechanism to suppress background noise interference, and a multi-scale decoupling detection head (MDH) that balances the detection accuracy of small and diffuse lesions. On a dataset containing six types of disease under various weather conditions, ToMASD achieves an average precision of 84.3%,.by a margin of 4.7% to 12.1% over thirteen mainstream models. The computational load is compressed to 7.1 GFLOPs. Through the introduction of a transfer learning paradigm, the pre-trained weights of the tomato disease detection model can be transferred to common bean and potato detection tasks. Through domain adaptation layers and adversarial feature decoupling strategies, the domain shift problem is overcome, achieving an average precision of 92.7% on the target crop test set. False detection rates in foggy and strong light conditions are controlled at 6.3% and 9.8%, respectively. This study achieves dual breakthroughs in terms of both high-precision detection in complex scenarios and the cross-crop generalization ability of lightweight models. It provides a new paradigm for universal agricultural disease monitoring systems that can be deployed at the edge.

Why it matches plant phenotyping methodsトマト葉の病斑・病害状態を画像から検出するモデルを開発し、複数モデル比較、悪条件評価、他作物への汎化検証を行っており、植物病害フェノタイピング手法が中心である。

abstractthis study presents a lightweight detection model called ToMASD.

Real-time jute leaf disease classification using an explainable lightweight CNN via a supervised and semi-supervised self-training approach. →

Introduction Timely detection of jute leaf diseases is vital for sustaining crop health and farmer livelihoods. Existing deep learning approaches often rely on large, annotated datasets, which are costly and time-consuming to produce. Methods and results To address this challenge, a lightweight convolutional neural network integrated with a semi-supervised learning self-training framework was proposed to enable accurate classification with minimal labeled data. The model combines modified depthwise separable convolutions, an enhanced squeeze-and-excite block, and a modified mobile inverted bottleneck convolution block, achieving strong representational power with only 2.24M parameters (8.54 MB). On a self-collected dataset of jute leaf images across three classes (Cescospora leaf spot, golden mosaic, and healthy leaf), the proposed model achieved a best accuracy of 98.95% under the supervised training with training, testing and validation split of 80:10:10. Remarkably, the model also attained a best accuracy of 97.89% in the semi-supervised learning (SSL) setting with only 10% labeled and 90% unlabeled data, demonstrating that near-supervised performance can be maintained while substantially reducing the dependency on costly labeled datasets. The application of explainable AI method such as Grad-CAM provided interpretable visualizations of diseased regions, and deployment as a Flask-based web application demonstrated practical, real-time usability in resource-constrained agricultural environments. Conclusion These results highlight the novelty of combining SSL with a lightweight CNN to deliver near-supervised performance, improved interpretability, and real-world applicability while substantially reducing the dependence on expert-labeled data.

Why it matches plant phenotyping methods葉画像から病害状態を推定する軽量CNNと半教師あり学習を開発・評価しており、植物病害フェノタイピング手法が研究の中心である。

abstracta lightweight convolutional neural network integrated with a semi-supervised learning self-training framework was proposed to enable accurate classification with minimal labeled data

A spontaneous keypoints connection algorithm for leafy plants skeletonization and phenotypes extraction. →

Introduction: Leaf phenotypes are key indicators of plant growth status. Existing deep learning-based leaf skeletonization typically requires extensive manual labeling, long training, and predefined keypoints, which limits scalability. We developed a training-free and label-free approach that connects spontaneously detected keypoints to generate leaf skeletons for leafy plants. Methods: The method comprises random seed-point generation and adaptive keypoint connection. For plants with random leaf morphology, we determine a threshold for the angle difference among any three consecutive adjacent points and iteratively identify keypoints within circular search neighborhoods to trace leaf skeletons. For plants with regular leaf morphology, we fit the skeleton trajectory by minimizing curvature. We validated the approach on vertical and front-view images of orchids (covering random and regular morphological cases) and extracted five phenotypic parameters from the resulting skeletons. Generalization was further assessed on a maize image dataset. Results: On orchid images, the proposed approach achieved an average curvature fitting error of 0.12 and an average leaf recall of 92%. Five orchid phenotypic parameters were accurately derived from the skeletons. The method also showed effective skeletonization on maize, indicating cross-species applicability. Discussion: By eliminating manual labels and training, this approach reduces annotation effort and computational overhead while enabling precise geometric phenotype calculation from skeleton-based keypoints. Its effectiveness on both randomly distributed and regularly shaped leafy plants suggests suitability for high-throughput plant phenotyping workflows.

Why it matches plant phenotyping methods葉の骨格化とフェノタイプ抽出のための画像解析手法を開発し、複数植物種で検証しているため、植物フェノタイピング手法が中心的である。

abstractWe developed a training-free and label-free approach that connects spontaneously detected keypoints to generate leaf skeletons for leafy plants.

Prediction of harvest-related traits in barley using high-throughput phenotyping data and machine learning. →

Developing crop varieties that maintain productivity under drought is essential for future food security. Here, we investigated the potential of time-resolved high-throughput phenotyping to predict harvest-related traits and identify drought-stressed plants. Six barley lines ( Hordeum vulgare ) were grown in a greenhouse environment with well-watered and drought treatments, and dynamically phenotyped using RGB, thermal infrared, chlorophyll fluorescence, and hyperspectral imaging sensors. A temporal phenomic classification model accurately distinguished between drought-treated and control plants, achieving high accuracy (classification accuracy ≥0.97) even when relying solely on predictors from the early drought response phase. Canopy temperature depression at the early stage and RGB-derived plant size estimates at the late stage emerged as key classification features. A temporal phenomic prediction model of harvest-related traits achieved particularly high mean R 2 values for total biomass dry weight (0.97) and total spike weight (0.93), with RGB plant size estimators emerging as important predictors. Importantly, prediction accuracy for these traits remained high (R 2 ≥ 0.84) even when restricted to early developmental phase data, including the stem elongation stage. Models trained on pooled drought and control data outperformed single-treatment models and maintained high predictive power across treatments. Together, these findings highlight the value of integrating high-throughput phenotyping with temporal modeling to enable earlier, more cost-effective selection of drought-resilient genotypes and demonstrate the broader potential of phenomics-driven strategies for accelerating crop improvement under stress-prone environments.

Why it matches plant phenotyping methodsRGB・熱赤外・蛍光・ハイパースペクトルによる時系列表現型取得と、収穫形質予測モデルの構築・評価が研究の中心であるため。

abstractdynamically phenotyped using RGB, thermal infrared, chlorophyll fluorescence, and hyperspectral imaging sensors

New-generation rice seed germination assessment: high efficiency and flexibility via SeedRuler web-based platform. →

Introduction The germination rate of rice seed is a critical indicator in agricultural research and production, directly influencing crop yield and quality. Traditional assessment methods based on manual visual inspection are often time-consuming, labor-intensive, and prone to subjectivity. Existing automated approaches, while helpful, typically suffer from limitations such as rigid germination standards, strict imaging requirements, and difficulties in handling the small size, dense arrangement, and variable radicle lengths of rice seeds. Methods To address these challenges, we present SeedRuler, a versatile, web-based application designed to improve the accuracy, efficiency, and usability of rice seed germination analysis. SeedRuler integrates three core components: SeedRuler-IP, a traditional image processing-based module; SeedRuler-YOLO, a deep learning model built on YOLOv5 for high-precision object detection; and SeedRuler-SAM, which leverages the Segment Anything Model (SAM) for fine-grained seed segmentation. A dataset of 1,200 rice seed images was collected and manually annotated to train and evaluate the system. An interactive module enables users to flexibly define germination standards based on specific experimental needs. Results SeedRuler-YOLO achieved a mean average precision (mAP) of 0.955 and a mean absolute error (MAE) of 0.110, demonstrating strong detection accuracy. Both SeedRuler-IP and SeedRuler-SAM support interactive germination standard customization, enhancing adaptability across diverse use cases. In addition, SeedRuler incorporates an automated seed size measurement function developed in our prior work, enabling efficient extraction of seed length and width from each image. The entire analysis pipeline is optimized for speed, delivering germination results in under 30 seconds per image. Conclusions SeedRuler overcomes key limitations of existing methods by combining classical image processing with advanced deep learning models, offering accurate, scalable, and user-friendly germination analysis. Its flexible standard-setting and automated measurement features further enhance usability for both researchers and agricultural practitioners. SeedRuler represents a significant advancement in rice seed phenotyping, supporting more informed decision-making in seed selection, breeding, and crop management.

Why it matches plant phenotyping methodsイネ種子の発芽状態と種子サイズを画像から抽出するウェブ型フェノタイピング手法を開発し、データセットで性能評価しているため、方法が研究の中心である。

abstractwe present SeedRuler, a versatile, web-based application designed to improve the accuracy, efficiency, and usability of rice seed germination analysis.

GCASSN: a graph convolutional attention synergistic segmentation network for 3D plant point cloud segmentation →

Plant phenotyping analysis serves as a cornerstone of agricultural research. 3D point clouds greatly improve the problem of overlapping and occlusion of leaves in two-dimensional images and have become a popular field of plant phenotyping research. The realization of faster and more effective plant point cloud segmentation is the basis and key to the subsequent analysis of plant phenotypic parameters. To balance lightweight design and segmentation precision, we propose a Graph Convolutional Attention Synergistic Segmentation Network (GCASSN) specifically for plant point cloud data. The framework mainly comprises (1) Trans-net, which normalizes input point clouds into canonical poses; (2) Graph Convolutional Attention Synergistic Module (GCASM), which integrates graph convolutional networks (GCNs) for local feature extraction and self-attention mechanisms to capture global contextual dependencies. Complementary advantages are realized. On plant 3D point cloud segmentation via the Plant3D and Phone4D datasets, the model achieves state-of-the-art performance with 95.46% mean accuracy and 90.41% mean intersection-over-union (mIoU), surpassing mainstream methods (PointNet, PointNet++, DGCNN, PCT, and Point Transformer). The computational efficiency is competitive, with the inference time and parameter quantity slightly exceeding that of the DGCNN. Without parameter tuning, it attains 85.47% mIoU and 82.9% mean class IoU on ShapeNet, demonstrating strong generalizability. The method proposed in this article can fully extract the local detail features and overall global features of plants, and efficiently and robustly complete the segmentation task of plant point clouds, laying a solid foundation for plant phenotype analysis. The code of the GCASSN can be found in https://github.com/fallovo/GCASSN.git.

Why it matches plant phenotyping methods植物3D点云のセグメンテーション手法を開発・ベンチマークし、植物表現型解析の基盤となるワークフローを提供しているため、方法が中心的である。

abstractwe propose a Graph Convolutional Attention Synergistic Segmentation Network (GCASSN) specifically for plant point cloud data.

ADQ-YOLOv8m: a precise detection model of sugarcane disease in complex environment. →

Introduction Current research on sugarcane disease identification primarily focuses on a limited number of typical diseases, often constrained by specific target groups or conditions. To address this, we propose an enhanced ADQ-YOLOv8m model based on the YOLOv8m framework, enabling precise detection of sugarcane diseases. Methods The detection head is modified to a Dynamic Head to enhance feature representation capabilities. Following the Detect module, we introduce the ATSS dynamic label assignment strategy and the QFocalLoss loss function to address issues such as class imbalance, thereby bolstering the model's feature representation capabilities. Results Experimental results demonstrate that ADQ-YOLOv8m outperforms nine other mainstream object detection models, achieving precision, recall, mAP50, mAP50-95, and F1 scores of 86.90%, 85.40%, 90.00%, 77.40%, and 86.00%, respectively. Discussion Finally, comprehensive evaluation of the ADQ-YOLOv8m model's performance is conducted using visual analysis of image predictions and cross-scenario adaptability testing. The experimental results indicate that the proposed model excels in multi-objective processing and demonstrates strong generalization capabilities, suitable for scenarios involving multiple objectives, multiple categories, and class imbalance. The detection method proposed exhibits excellent detection performance and potential, providing robust support for the development of intelligent sugarcane cultivation and disease control.

Why it matches plant phenotyping methodsサトウキビ病害の画像から病害状態を推定する検出モデルを開発・評価しており、植物の病害表現型取得が中心的な方法論的貢献である。

abstractwe propose an enhanced ADQ-YOLOv8m model based on the YOLOv8m framework, enabling precise detection of sugarcane diseases.

A lightweight deep convolutional neural network development for soybean leaf disease recognition. →

Soybean is one of the world's major oil-bearing crops and occupies an important role in the daily diet of human beings. However, the frequent occurrence of soybean leaf diseases caused serious threats to its yield and quality during soybean cultivation. Rapid identification of soybean leaf diseases could provide a better solution for efficient control and subsequent precision application. In this study, a lightweight deep convolutional neural network (CNN) based on multiscale feature extraction fusion (MFEF) and combined with a dense connectivity (DC) network (MFEF-DCNet) was proposed for soybean leaf disease identification. In MFEF-DCNet, a multiscale feature extraction fusion (MFEF) module for soybean leaves was constructed by utilizing a convolutional attention module and depth-separable convolution to improve the model feature extraction capability. Multiscale features are fused by using dense connections (DC) in the backbone network to improve the model generalization capability. Experiments were implemented on eight distinct disease and deficiency classes of soybean images (including bacterial blight, cercospora leaf blight, downy mildew, frogeye leaf spot, healthy, potassium deficiency, soybean rust, and target spot) using the proposed network. The results showed that the MFEF-DCNet had an accuracy of 0.9470, an average precision of 0.9510, an average recall of 0.9480, and an F1-score of 0.9490 for soybean leaf disease identification. And MFEF-DCNet had certain performance advantages in terms of classification accuracy, convergence speed and other effects compared with VGG16, ResNet50, DenseNet201, EfficientNetB0, Xception and MobileNetV3_small models. In addition, the accuracy of the MFEF-DCNet model in recognizing soybean diseases in local data was 0.9024, which indicated that the MFEF-DCNet model had favorable application in practical applications. The proposed model and experience in this study could provide useful inspiration for automated disease identification in soybean and other crops.

Why it matches plant phenotyping methods大豆葉の病害状態を画像から認識するCNNを開発・評価しており、植物病害フェノタイピング手法が研究の中心である。

abstracta lightweight deep convolutional neural network (CNN) based on multiscale feature extraction fusion (MFEF) and combined with a dense connectivity (DC) network (MFEF-DCNet) was proposed for soybean leaf disease identification.

Enhancing yield prediction from plot-level satellite imagery through genotype and environment feature disentanglement. →

Accurately predicting yield during the growing season enables improved crop management and better resource allocation for both breeders and growers. Existing yield prediction models for an entire field or individual plots are based on satellite-derived vegetation indices (VIs) and widely used machine learning-based feature extraction models, including principal component analysis (PCA) and autoencoders (AE). Here, we significantly enhance pre-harvest yield prediction at plot-scale using Compositional Autoencoders (CAE) - a deep-learning-based feature extraction approach designed to disentangle genotype (G) and environment (E) features - on high-resolution, plot-level satellite imagery. Our approach uses a dataset of approximately 4,000 satellite images collected from replicated plots of 84 hybrid maize varieties grown at five distinct locations across the U.S. Corn Belt. By deploying the CAE model, we improve the separation of genotype and environment effects, enabling more accurate incorporation of genotype-by-environment (GxE) interactions for downstream prediction tasks. Results show that the CAE-based features improve early-stage yield predictions by up to 10% compared to traditional autoencoder-based features and outperform vegetation indices (VIs) by 9% across various growth stages. The CAE model also excels in separating environmental factors, achieving a high silhouette score of 0.919, indicating effective clustering of environmental features. Moreover, the CAE consistently outperforms standard models in unseen environments and unseen genotypes yield predictions, demonstrating strong generalizability. This study demonstrates the value of disentangling G and E effects for providing more accurate and early yield predictions that support informed decision-making in precision agriculture and plant breeding.

Why it matches plant phenotyping methods作物プロットの収量という植物形質を、衛星画像から推定する深層学習特徴抽出法を開発・評価しており、表現型取得・推定手法が研究の中心である。

abstractHere, we significantly enhance pre-harvest yield prediction at plot-scale using Compositional Autoencoders (CAE) - a deep-learning-based feature extraction approach designed to disentangle genotype (G) and environment (E) features - on high-resolution, plot-level satellite imagery.

Harnessing artificial intelligence for sustainable rice leaf disease classification. →

Introduction Agriculture underpins global food security by providing food, raw materials, and livelihoods, contributing 4% to global GDP and up to 25% in rural areas. Rice, a staple for more than half of the world's population, is nutritionally vital but highly vulnerable to diseases such as Hispa, leaf blast, and brown spots, which significantly reduce yield and quality. Achieving Sustainable Development Goal (SDG) 2 requires innovative approaches to mitigate these threats. Artificial intelligence (AI), particularly computer vision and machine learning, offers promising tools for early disease detection. Methods This study developed a convolutional neural network (CNN)-based model for rice leaf disease detection and classification. A publicly available dataset containing 3,355 labeled images across four categories-Brown Spot, Leaf Blast, Hispa, and Healthy leaves-was used to train and evaluate the model. To improve classification accuracy, the CNN was enhanced with spatial and channel attention mechanisms, enabling it to focus on the most discriminative image regions. The system was designed for modular deployment, allowing lightweight, real-time implementation on edge devices. Results The enhanced CNN achieved high accuracy and robust performance metrics across all disease categories. Attention mechanisms significantly improved precision in identifying subtle disease patterns. The lightweight design ensured efficient operation on edge devices, demonstrating feasibility for real-world agricultural applications. Discussion and conclusion The proposed AI-driven system provides reliable and scalable rice leaf disease detection, supporting timely intervention to reduce yield loss. By strengthening rice production and promoting sustainable practices, the model contributes to SDG 2 by advancing global food security. This research highlights AI's transformative role in agriculture, fostering mechanization, ecological stability, and resilience in food systems.

Why it matches plant phenotyping methodsイネ葉の病徴を画像から分類するCNN手法を開発・評価しており、植物の病害状態を直接推定するフェノタイピング手法が研究の中心である。

abstractThis study developed a convolutional neural network (CNN)-based model for rice leaf disease detection and classification.

YOLOv8-FDA: lightweight wheat ear detection and counting in drone images based on improved YOLOv8. →

Introduction Wheat is a vital global staple crop, where accurate ear detection and counting are essential for yield prediction and field management. However, the complexity of field environments poses significant challenges to achieving lightweight yet high-precision detection. Methods This study proposes YOLOv8-FDA, a lightweight detection and counting method based on YOLOv8. The approach integrates RFAConv for enhanced feature extraction, DySample for efficient multi-scale upsampling, HWD for compressed and accelerated model training, and the SDL loss for improved bounding box regression. Results Experimental results on the GWHD dataset show that YOLOv8-FDA achieves a precision of 86.3%, recall of 77.5%, and mAP@0.5 of 84.9%, outperforming the original YOLOv8n by significant margins. The model size is 2.96MB with a computational cost of 8.3 GFLOPs, and it operates at 19.2 FPS, enabling real-time counting with over 97.5% accuracy using cross-row segmentation. Discussion The proposed YOLOv8-FDA model demonstrates strong detection performance, lightweight characteristics, and efficient real-time capability, indicating its high practicality and suitability for deployment in real-world agricultural applications.

Why it matches plant phenotyping methodsコムギ穂の検出・計数という植物器官形質の画像ベース抽出法をYOLOv8改良モデルとして開発し、データセット上で性能検証しているため、方法が研究の中心である。

abstractThis study proposes YOLOv8-FDA, a lightweight detection and counting method based on YOLOv8.

Deep learning driven, image-based phenotyping of seed processing efficiency in sainfoin ( Onobrychis viciifolia ). →

Introduction: spp.) is a perennial legume traditionally cultivated as a forage crop and is now emerging as a promising candidate for development as a perennial grain legume. Despite its potential, no research has addressed the breeding of sainfoin varieties with superior grain processing properties. Methods: We conducted a multifactorial experiment to evaluate the depodding and dehulling efficiency of five commercially available sainfoin varieties. Seeds were processed using two different methods (belt thresher and impact dehuller) across five sample sizes. A pre-trained Faster R-CNN (Region-based Convolutional Neural Network) object detection model was fine-tuned to identify intact pods, whole seeds, and split seeds from images of the processed mixtures. These predictions were used to calculate processing efficiency (PE) for each variety. A comprehensive power analysis was performed to determine the minimum sample size of sainfoin pods required to detect differences in PE with high statistical power. Results: We observed strong varietal differences in PE, as well as clear effects of the processing method. Belt threshing produced mixtures with more intact pods, while the impact dehuller generated a higher proportion of split seeds. Increasing sample size led to more intact pods across all varieties and methods, and notably decreased seed proportion in belt-threshed samples. Statistical modeling combined with object detection outputs revealed that a minimum of 2 g of pods is required to reliably detect an absolute proportional difference of 0.25 in PE between two breeding lines with 80% power. Discussion: Our findings demonstrate that sainfoin varieties differ significantly in processing efficiency and that processing outcomes depend strongly on both method and sample size. Integrating deep learning-based phenotyping with robust statistical design enables efficient evaluation of processing traits and provides actionable guidelines for breeding programs. While deep learning models offer powerful, cost-effective tools for plant phenotyping, their outputs must be paired with rigorous statistical design to yield reliable and actionable insights for crop improvement.

Why it matches plant phenotyping methods画像からポッド・種子を検出し、処理効率という植物由来形質を算出する深層学習ベースの表現型解析が研究の中心であるため。

titleDeep learning driven, image-based phenotyping of seed processing efficiency in sainfoin

A novel efficient eggplant disease detection method with multi-scale learning and edge feature enhancement. →

In the context of the rapid development of smart agriculture, the detection of crop diseases remains a critical and challenging task. The diversity in eggplant disease scales, disease edge features, and the complexity of planting backgrounds significantly impact disease detection effectiveness. To address these challenges, we propose an eggplant disease detection network with edge feature enhancement based on multi-scale learning. The overall network adopts a "backbone-neck-head" architecture: the backbone extracts features, the neck performs feature fusion, and a three-scale detection head produces the final predictions. First, we designed the Multi-scale Edge Information Enhance (CSP-MSEIE) module to extract features from different disease scales and highlight edge information to obtain richer target features. Second, the Multi-source Interaction Module (MSIM) and Dynamic Interpolation Interaction Module (DIIM) sub-modules were designed further to enhance the model's capacity for multi-scale feature representation. By leveraging dynamic interpolation and feature fusion strategies, these sub-modules significantly improved the model's ability to detect targets in complex backgrounds. Then, leveraging these sub-modules, we designed the Multi-scale Context Reconstruction Pyramid Network (MCRPN) to facilitate spatial feature reconstruction and hierarchical context extraction. This framework efficiently combines feature information across multiple levels, strengthening the model's ability to capture and utilize contextual details. Finally, we validated the effectiveness of the proposed model on two disease datasets. It is noteworthy that on the eggplant disease data, the proposed disease detection model achieved improvements of 4.7% and 7.2% in mAP50 and mAP50-95 metrics, respectively, and the model's frames per second (FPS) reached 270.5. This detection network provides an effective solution for the efficient detection of crop diseases.

Why it matches plant phenotyping methods植物病害の画像ベース検出ネットワークを開発し、病害状態の推定をデータセットで検証しており、表現型取得・抽出手法が中心である。

abstractwe propose an eggplant disease detection network with edge feature enhancement based on multi-scale learning.

Non-destructive estimation of SPAD and biomass in Lamiophlomis rotata using hyperspectral imaging and deep learning with DRSA-CARS feature selection. →

Introduction Monitoring the growth status and aboveground biomass of wild and cultivated medicinal herbs remains a persistent challenge in precision agriculture. Methods In this study, we developed machine learning and deep learning models to estimate SPAD values and biomass of Lamiophlomis rotata (Benth.). The models used hyperspectral data and time-series phenotypic traits from 508 samples collected across different altitudes. Regions of interest (ROIs) were manually defined from plant contours. The corresponding mean spectral profiles were then preprocessed. To improve feature selection, we proposed a Dynamic Reptile Search Algorithm-enhanced CARS (DRSA-CARS) method. This method integrates a dynamic behavioral strategy into the CARS framework to identify informative spectral bands. Vegetation indices (VIs) and gray-level co-occurrence matrix (GLCM)-based texture parameters were extracted and combined with spectral features to construct the PLSR, SVR, FNN, and CNN models. Results Compared to CARS, the DRSA-CARS method reduced feature dimensionality by up to 75.7% for SPAD and 29.2% for biomass, while improving prediction accuracy ( R ²) by 24.4% and 34.7%, respectively. Among all models, the FNN achieved the highest performance, with R ² values of 0.7732 (training) and 0.7502 (testing) for SPAD and 0.8260 and 0.7933 for biomass. Feature fusion further improved predictive accuracy by 11% for SPAD and 30% for biomass compared to models based on individual feature types. Discussion These results demonstrate that coupling DRSA-CARS-based feature selection with deep learning provides a robust, non-destructive approach for evaluating plant growth status. This framework highlights the potential of hyperspectral imaging as a rapid, reliable, non-invasive tool for precision cultivation of medicinal herbs.

Why it matches plant phenotyping methodsハイパースペクトル画像からSPAD値とバイオマスという植物形質を非破壊推定し、特徴選択法と深層学習モデルを開発・評価しており、フェノタイピング手法が中心である。

abstractwe developed machine learning and deep learning models to estimate SPAD values and biomass of Lamiophlomis rotata (Benth.).

YOLO-LF: application of multi-scale information fusion and small target detection in agricultural disease detection. →

With the increasing threat of agricultural diseases to crop production, traditional manual detection methods are inefficient and highly susceptible to environmental factors, making an efficient and automated disease detection method urgently needed. Existing deep learning models still face challenges in detecting small targets and recognizing multi-scale lesions in complex backgrounds, particularly in terms of multi-feature fusion. To address these issues, this paper proposes an improved YOLO-LF model by introducing modules such as CSPPA (Cross-Stage Partial with Pyramid Attention), SEA (SeaFormer Attention), and LGCK (Local Gaussian Convolution Kernel), aiming to improve the accuracy and efficiency of small target disease detection. Specifically, the CSPPA module enhances multi-scale feature fusion, the SEA module strengthens the attention mechanism for contextual and local information to improve detection accuracy, and the LGCK module increases the model's sensitivity to small lesion areas. Experimental results show that the proposed YOLO-LF model achieves significant performance improvements on the Plant Pathology 2020 - FGVC7 and Plant Pathology 2021 - FGVC8 datasets, particularly in mAP@0.5% and mAP@0.5-0.95%, outperforming existing mainstream models. These results indicate that the proposed method effectively handles complex backgrounds and small target detection tasks in agricultural disease detection, demonstrating high practical value.

Why it matches plant phenotyping methods植物病斑という植物の病害状態を画像から検出・推定する深層学習モデルを開発し、複数データセットで性能評価しているため、植物フェノタイピング手法が中心です。

abstractthis paper proposes an improved YOLO-LF model by introducing modules such as CSPPA (Cross-Stage Partial with Pyramid Attention), SEA (SeaFormer Attention), and LGCK (Local Gaussian Convolution Kernel), aiming to improve the accuracy and efficiency of small target disease detection.

MFDN: an efficient detection method for Alstroemeria Genus flowers based on multi-scale feature fusion. →

As an ornamental plant, Alstroemeria Genus Morado holds great significance in precision agriculture for the automatic detection and classification of its flower maturity. However, due to its diverse morphologies, complex growth environments, and factors such as occlusion and lighting changes, related tasks face numerous challenges, and research in this area is relatively scarce. This study proposes a deep - learning - based object detection framework, the Morado Flower Detection Network (MFDN), which consists of two parts: a backbone network and a head network. Novel modules such as C3k2_PPA are introduced. Through multi - branch fusion and the attention mechanism, the ability to detect small targets is enhanced. The head network uses the CARAFE module for upsampling, combines features through Concat, accelerates processing with the optimized C2f module, and finally achieves precise detection and classification through the Detect module. In the comparative experiment on the morado_5may dataset, MFDN performs outstandingly in indicators such as Precision, Recall, and F1 - score. The mean Average Precision (mAP) of MFDN is 1.3% - 5.8% higher than that of YOLO - series models. It has strong generalization ability and is expected to contribute to improving the efficiency and automation level of agricultural production.

Why it matches plant phenotyping methodsアルストロメリア花の成熟度を画像から検出・分類する深層学習手法を開発し、データセット上で比較評価しており、植物表現型取得が中心である。

abstractThis study proposes a deep - learning - based object detection framework, the Morado Flower Detection Network (MFDN)

A novel transformer using dynamic range-enhanced discrete cosine transform for detecting bean leaf diseases. →

Introduction Early detection of diseases on bean leaves is essential for preventing declines in agricultural productivity and mitigating broader agricultural challenges. However, some bean leaf diseases are difficult to detect even with the human eye, posing significant challenges for machine learning methods that rely on precise feature extraction. Methods We propose a novel approach, DCT-Transformers, which combines a preprocessing technique, dynamic range enhanced discrete cosine transform (DRE-DCT) with Transformer-based models. The DRE-DCT method enhances the dynamic range of input images by extracting high-frequency components and subtle details that are typically imperceptible while preserving overall image quality. Transformer models were then used to classify bean leaf images before and after applying this preprocessing step. Results Experimental evaluations demonstrate that the proposed DCT-Transformers method achieved a classification accuracy of 99.56% (precision: 0.9916, recall: 0.9912, F1-score: 0.9912) when using preprocessed images, compared to 95.92% when using non-preprocessed images. Moreover, the method outperformed state-of-the-art approaches (all below 94%) and similar studies (all below 98.5%). Discussion These findings indicate that enhancing feature extraction through DRE-DCT significantly improves disease classification performance. The proposed method offers an efficient solution for early disease detection in agriculture, contributing to improved disease management strategies and supporting food security initiatives.

Why it matches plant phenotyping methods豆の葉画像から病害状態を推定する画像・計算手法を提案し、前処理とTransformerの性能を比較評価しており、植物フェノタイピング手法が研究の中心である。

abstractWe propose a novel approach, DCT-Transformers, which combines a preprocessing technique, dynamic range enhanced discrete cosine transform (DRE-DCT) with Transformer-based models.

LCAMNet: a lightweight model for apple leaf disease classification in natural environments. →

Apple leaf diseases severely affect the quality and yield of apples, and accurate classification is crucial for reducing losses. However, in natural environments, the similarity between backgrounds and lesion areas makes it difficult for existing models to balance lightweight design and high accuracy, limiting their practical applications. In order to resolve the aforementioned problem, this paper introduces a lightweight converged attention multi-branch network named LCAMNet. The network integrates depthwise separable convolutions and structural re-parameterization techniques to achieve efficient modeling. To avoid feature loss caused by single downsampling operations, a dual-branch downsampling module is designed. A multi-scale structure is introduced to enhance lesion feature diversity representation. An improved triplet attention mechanism is utilized to better capture deep lesion features. Furthermore, a dataset named SCEBD is constructed, containing multiple common disease types and interference factors under natural environments, realistically reflecting orchard conditions. Experimental results show that LCAMNet achieves 92.60% accuracy on the SCEBD and 95.31% on a public dataset, with only 0.03 GFLOPs and 1.30M parameters. The model maintains high accuracy while remaining lightweight, enabling effective apple leaf disease classification in natural environments on devices with limited resources.

Why it matches plant phenotyping methodsリンゴ葉の病徴を画像から分類する軽量モデルを開発し、自然環境データセットを構築・評価しており、植物病害状態の画像ベース表現型推定が中心である。

abstractthis paper introduces a lightweight converged attention multi-branch network named LCAMNet.

LiSA-MobileNetV2: an extremely lightweight deep learning model with Swish activation and attention mechanism for accurate rice disease classification. →

In the context of intelligent agriculture in China, rapid and accurate identification of crop diseases is essential for ensuring food security and improving crop yield. Although lightweight convolutional neural networks (CNNs) are widely adopted for plant disease recognition due to their computational efficiency, they often suffer from limited feature representation and classification accuracy. To address these challenges, we propose LiSA-MobileNetV2, an improved MobileNetV2-based model designed for rice disease classification. First, we restructure the inverted residual module to simplify the network architecture, achieving a test accuracy of 92.32%, representing a 2.41% improvement over the original MovileNetV2 (89.91%). This indicate that a more lightweight network can enhance feature representation in specific disease recognition. Second, integrating the Swish activation function further improves accuracy to 94.04% by enhancing the model's non-linear feature learning. Finally, the addition of a squeeze-and-excitation attention mechanism raises accuracy to 95.68%, representing a 5.77% improvement over the original model. Importantly, the parameter size and FLOPs are reduced by 74.69% and 48.18%, respectively, maintaining strong computational efficiency. These results demonstrate that combining structural simplification, advanced activation, and efficient attention mechanisms significantly improves CNN performance. LiSA-MobileNetV2 provides a high-accuracy, resource-efficient solution for real-time rice disease detection in smart farming systems.

Why it matches plant phenotyping methodsイネ病害の症状を画像から分類する深層学習モデルを開発・評価しており、植物の病害状態の取得・推定が中心的な方法論的貢献である。

abstractwe propose LiSA-MobileNetV2, an improved MobileNetV2-based model designed for rice disease classification.

Enhancing leaf disease classification using GAT-GCN hybrid model. →

Agriculture plays a critical role in the global economy, providing livelihoods and ensuring food security for billions. Progress in agricultural techniques has helped boost crop yield, along with a growing need for precise disease monitoring solutions. This requires accurate, efficient, and timely disease detection methods. The research presented in this paper addresses this need by analyzing a hybrid model built using Graph Attention Network (GAT) and Graph Convolution Network (GCN) models. The integration of these models has witnessed a notable improvement in the accuracy of leaf disease classification. GCN has been widely used for learning from graph-structured data, and GAT enhances this by incorporating attention mechanisms to focus on the most important neighbors. The methodology incorporates superpixel segmentation for efficient feature extraction, partitioning images into meaningful, homogeneous regions that better capture localized features. The robustness of the model is further enhanced by the edge augmentation technique. The edge augmentation technique in the context of graph has introduced a significant degree of generalization in the detection capabilities of the model as analyzed on apple, potato, and sugarcane leaves. To further optimize training, weight initialization techniques are applied. The hybrid model is evaluated against the individual performance of the GCN and GAT models and the hybrid model achieved a precision of 0.9822, recall of 0.9818, and F1-score of 0.9818 in apple leaf disease classification, a precision of 0.9746, recall of 0.9744, and F1-score of 0.9743 in potato leaf disease classification, and a precision of 0.8801, recall of 0.8801, and F1-score of 0.8799 in sugarcane leaf disease classification. The results indicate that the model is effective and consistent in identifying leaf diseases in plants.

Why it matches plant phenotyping methods植物葉画像から病害状態を分類するGAT-GCNモデルを開発・比較評価しており、病害表現型の抽出手法が中心である。

abstractThe methodology incorporates superpixel segmentation for efficient feature extraction, partitioning images into meaningful, homogeneous regions that better capture localized features.

BiSeNeXt: a yam leaf and disease segmentation method based on an improved BiSeNetV2 in complex scenes. →

Introduction Yam is an important medicinal and edible crop, but its quality and yield are greatly affected by leaf diseases. Currently, research on yam leaf disease segmentation remains unexplored. Challenges like leaf overlapping, uneven lighting and irregular disease spots in complex environments limit segmentation accuracy. Methods To address these challenges, this paper introduces the first yam leaf disease segmentation dataset and proposes BiSeNeXt, an enhanced method based on BiSeNetV2. Firstly, dynamic feature extraction block (DFEB) enhances the precision of leaf and disease edge pixels and reduces lesion omission through dynamic receptive-field convolution (DRFConv) and pixel shuffle (PixelShuffle) downsampling. Secondly, efficient asymmetric multi-scale attention (EAMA) effectively alleviates the problem of lesion adhesion by combining asymmetric convolution with a multi-scale parallel structure. Finally, PointRefine decoder adaptively selects uncertain points in the image predictions and refines them point-by-point, producing accurate segmentation of leaves and spots. Results Experimental results indicated that the approach achieved a 97.04% intersection over union (IoU) for leaf segmentation and an 84.75% IoU for disease segmentation. Compared to DeepLabV3+, the proposed method improves the IoU of leaf and disease segmentation by 2.22% and 5.58%, respectively. Additionally, the FLOPs and total number of parameters of the proposed method require only 11.81% and 7.81% of DeepLabV3+, respectively. Discussion Therefore, the proposed method can efficiently and accurately extract yam leaf spots in complex scenes, providing a solid foundation for analyzing yam leaves and diseases.

Why it matches plant phenotyping methodsヤム葉と病斑を画像から分割・抽出する手法とデータセットを開発し、性能比較まで行っており、植物の病害状態を取得する方法が中心である。

abstractthis paper introduces the first yam leaf disease segmentation dataset and proposes BiSeNeXt, an enhanced method based on BiSeNetV2.

Correlative X-ray imaging to reveal the dissolution of nanoparticles and nutrient transport in plant foliar fertilization →

The integration of nanotechnology in agriculture allows for more precise nutrient delivery through nanoparticles (NPs), particularly via foliar application. To mature this technology for enhancing fertilizer efficiency, it is essential to shed new light on the transport and dissolution of NPs in plants. Available analytical methods struggle to address this challenge in a direct manner. We introduce correlative X-ray imaging as a novel analytical tool capable of tracking NP pathways, dissolution and hence nutrient release in plants. By utilizing three complementary X-ray techniques, we offer a unique insight into the plant processes associated with foliar fertilization. We demonstrate that small-angle X-ray scattering enables the characterization of NP size and concentration, while X-ray fluorescence imaging, maps the distribution of elements within the sample. Finally, micro-computed tomography integrates these findings into a complete three-dimensional digital representation of the plant’s microstructure, revealing regions of apparent densification associated with NP accumulation. Using freeze-dried barley plants infiltrated with nano-hydroxyapatite (nHAP), we observed rapid dissolution of NPs, and we are able to associate time and space attributes to the translocation process of nutrients up to three days following foliar application of NPs. With the first pilot study of applying correlative X-ray imaging to live plants, we sought to indicate the potential of this new analytical approach for future nano-enabled agricultural research.

Why it matches plant phenotyping methods植物内のナノ粒子経路・溶解・栄養輸送を可視化する相関X線イメージング手法の導入と実証が中心であり、植物の状態・生理過程を測定する方法論的研究である。

abstractWe introduce correlative X-ray imaging as a novel analytical tool capable of tracking NP pathways, dissolution and hence nutrient release in plants.

Plant disease classification in the wild using vision transformers and mixture of experts. →

Plant disease classification using deep learning techniques has shown promising results, especially when models are trained on high-quality images. However, these models often suffer from a significant drop in their accuracies when tested in real-world agricultural settings. In the wild, models encounter images that are significantly different from the training data in aspects like lighting conditions, capturing conditions, image resolution, and the severity of disease. This discrepancy between the training images and images in-the-wild conditions poses a major challenge for deploying these models in agricultural settings. In this paper, we present a novel approach to address this issue by combining a Vision Transformer backbone with a Mixture of Experts, where multiple expert models are trained to specialize in different aspects of the input data, and a gating mechanism is implemented to select the most relevant experts for each input. The use of Mixture of Experts allows the model to dynamically allocate specialized experts to different types of input data, improving model performance across diverse image conditions. The approach significantly improves performance on diverse datasets that contain a range of image capturing conditions and disease severities. Furthermore, the model incorporates entropy regularization and orthogonal regularization, aiming to enhance the robustness and generalization capabilities. Experimental results demonstrate that the proposed model achieved a 20% improvement in accuracy compared to Vision Transformer (ViT). Furthermore, it demonstrated a 68% accuracy on cross-domain datasets like PlantVillage to PlantDoc, surpassing baseline models such as InceptionV3 and EfficientNet. This highlights the potential of our model for effective deployment in dynamic agricultural environments.

Why it matches plant phenotyping methods野外画像から植物病害の状態・重症度を推定する分類手法を開発し、異なるデータセットや撮影条件で性能検証しているため、植物フェノタイピング手法が中心です。

abstractIn this paper, we present a novel approach to address this issue by combining a Vision Transformer backbone with a Mixture of Experts

Study on the germination rate of maize seeds based on improved YOLOv8n model. →

The germination potential of corn seeds, a key index for assessing their quality and directly associated with the ultimate corn yield, is currently defined in a way that cannot effectively portray the seed germination rate, and the prevalent measurement methods are traditional, consuming substantial process resources. To tackle these issues, this paper employs a public corn seed germination dataset, adds noise to it to simulate real - world production conditions, and ultimately acquires a dataset comprising 8148 images. It then proposes an enhanced YOLOv8 target detection model, EBS - YOLOv8, for detecting corn seed germination. Specifically, the ECA lightweight attention mechanism is introduced to decrease small - target feature loss, assist in accurate target recognition, and remove redundant features; simultaneously, the P2BiFPN multiscale feature fusion technique is utilized to boost the detection ability for small targets; furthermore, the ScConv convolution is adopted to enhance the feature - extraction capacity and improve detection accuracy. Combined with the improved model, this paper also proposed a mathematical modeling algorithmnew method for measuring seed germination potential and observing seed germination rate. The results indicate that the proposed model attains a mean average precision at 50% Intersection over Union (mAP50) value of 98.9%, a mean average precision in the range of 50% - 95% Intersection over Union (mAP50 - 95) value of 95.8%, an accuracy of 96.7%, and a recall of 96.3%. In comparison with the original model, the mAP50 has increased by 0.9% and the mAP50 - 95 value has witnessed a 3.7% increment. The experiments have demonstrated that the research method for germination potential put forward in this paper can effectively depict the rate variation of seeds during the germination process, thus offering a novel perspective for future research on seed germination potential.

Why it matches plant phenotyping methodsトウモロコシ種子の発芽状態・発芽率を画像から検出・定量する改良YOLOv8モデルと数理測定法を開発し、データセット上で性能検証しているため、植物フェノタイピング手法が中心である。

abstractIt then proposes an enhanced YOLOv8 target detection model, EBS - YOLOv8, for detecting corn seed germination.

Nighttime fluorescence phenotyping reduces environmental variability for photosynthetic traits and enables the identification of candidate loci in maize. →

Introduction: Photosynthesis is fundamental to agricultural productivity, but its relatively low light-to-biomass conversion efficiency represents an opportunity for enhancement. High-throughput phenotyping is crucial for unraveling the genetic basis of variation in photosynthetic activity. However, the heritability of chlorophyll fluorescence parameters measured during the day is often low as a result of high levels of variation introduced by environmental fluctuations. Methods: To address these limitations, we measured fluorescence phenotypes at night, leveraging natural dark adaptation to minimize environmental noise. Results: Night measurement significantly increased the heritability of fluorescence traits compared to daytime measurements, with the maximum quantum yield of photosystem II (Fv/Fm) showing an increase in heritability from 0.32 to 0.72. Genome-wide association studies (GWAS) conducted using three photosynthetic fluorescence traits measured at night across two growing seasons identified several significant single nucleotide polymorphisms (SNPs). Notably, two candidate genes near SNPs linked to multiple fluorescence traits, Zm00001eb271820 and Zm00001eb012130, have known roles in photosynthesis regulation. Four of the significant signal nucleotide polymorphisms identified in GWAS conducted using nighttime collected data also exhibited statistically significant associations with the same phenotypes during the day. In a majority of other cases, direction of effect was consistent but greater variance in day measured data relative to night measured data resulted in the differences not being statistically significant. Discussion: These results highlight the effectiveness of phenotyping photosynthetic traits at night in reducing environmental noise and enhancing the discovery of genomic intervals related to photosynthesis. While nighttime data collection may not be applicable for all photosynthetic traits, it offers a promising avenue for advancing our understanding of the genetic variation of photosynthesis in modern crop species.

Why it matches plant phenotyping methods夜間の蛍光測定による光合成形質フェノタイピングを開発・評価し、昼間測定との比較で環境変動低減と遺伝率向上を検証しているため、手法が研究の中心である。

abstractHigh-throughput phenotyping is crucial for unraveling the genetic basis of variation in photosynthetic activity.

DeepD&Cchl: an AI tool for automated 3D single-cell chloroplast detection, counting, and cell type clustering →

Chloroplast density in cells varies among different types of cells and plants. In current single-cell spatiotemporal analysis, the automatic detection and quantification of chloroplasts at the single-cell level is crucial. We developed DeepD&Cchl (Deep-learning-based Detecting-and-Counting-chloroplasts), an AI tool for single-cell chloroplast detection and cell-type clustering. It utilizes You-Only-Look-Once (YOLO), a real-time detection algorithm, for accurate and efficient performance. DeepD&Cchl has been proved to identify chloroplasts in plant cells across various imaging types, including light microscopy, electron microscopy, and fluorescence microscopy. Integrated with an Intersection Over Union (IOU) module, DeepD&Cchl precisely counts chloroplasts in single- or multi-layered images, while eliminating double-counting errors. Furthermore, when combined with Cellpose, a single-cell segmentation tool, DeepD&Cchl enhances its effectiveness at the single-cell level. By counting chloroplasts within individual cells, it supports cell-type-specific clustering based on chloroplast number versus cell size, offering valuable morphological insights for single-cell studies. In summary, DeepD&Cchl is a significant advancement in plant cell analysis. It offers accuracy and efficiency in chloroplast identification, counting and cell-type classification, providing a useful tool for plant research.

Why it matches plant phenotyping methods植物細胞画像から葉緑体を検出・計数し、細胞型をクラスタリングするAIツールの開発が中心であり、植物の形態的状態を定量化するフェノタイピング手法に該当する。

abstractWe developed DeepD&Cchl (Deep-learning-based Detecting-and-Counting-chloroplasts), an AI tool for single-cell chloroplast detection and cell-type clustering.

Automated dynamic phenotyping of whole oilseed rape ( Brassica napus ) plants from images collected under controlled conditions. →

Introduction Recent advancements in sensor technologies have enabled collection of many large, high-resolution plant images datasets that could be used to non-destructively explore the relationships between genetics, environment and management factors on phenotype or the physical traits exhibited by plants. The phenotype data captured in these datasets could then be integrated into models of plant development and crop yield to more accurately predict how plants may grow as a result of changing management practices and climate conditions, better ensuring future food security. However, automated methods capable of reliably and efficiently extracting meaningful measurements of individual plant components (e.g. leaves, flowers, pods) from imagery of whole plants are currently lacking. In this study, we explore interdisciplinary application of MapReader, a computer vision pipeline for annotating and classifying patches of larger images that was originally developed for semantic exploration of historical maps, to time-series images of whole oilseed rape (Brassica napus) plants. Methods Models were trained to classify five plant structures in patches derived from whole plant images (branches, leaves, pods, flower buds and flowers), as well as background patches. Three modelling methods are compared: (i) 6-label multi-class classification, (ii) a chain of binary classifiers approach, and (iii) an approach combining binary classification of plant and background patches, followed by 5-label multi-class classification of plant structures. Results A combined plant/background binarization and 5-label multi-class modelling approach using a ‘resnext50d_4s2x40d’ model architecture for both the binary classification and multi-class classification components was found to produce the most accurate patch classification for whole B. napus plant images (macro-averaged F1-score = 88.50, weighted average F1-score = 97.71). This combined binary and 5-label multi-class classification approach demonstrate similar performance to the top-performing MapReader ‘railspace’ classification model. Discussion This highlights the potential applicability of the MapReader model framework to images data from across scientific and humanities domains, and the flexibility it provides in creating pipelines with different modelling approaches. The pipeline for dynamic plant phenotyping from whole plant images developed in this study could potentially be applied to imagery from varied laboratory conditions, and to images datasets of other plants of both agricultural and conservation concern.

Why it matches plant phenotyping methods植物全体画像から葉・花・莢などの構造を自動抽出・分類する動的フェノタイピング手法を開発・評価しており、方法が研究の中心である。

abstractautomated methods capable of reliably and efficiently extracting meaningful measurements of individual plant components (e.g. leaves, flowers, pods) from imagery of whole plants are currently lacking.

Plant photosynthesis in basil (C3) and maize (C4) under different light conditions as basis of an AI-based model for PAM fluorescence/gas-exchange correlation →

Photosynthetic activity can be monitored using pulse amplitude modulated (PAM) fluorescence or gas exchange. While PAM provides insight into the light-dependent reactions, gas exchange reflects CO 2 fixation and water balance. Accurate, non-invasive prediction of photosynthetic performance under varying conditions is highly relevant for phenotyping and stress diagnostics. Despite their physiological link, data from both methods do not always correlate. To systematically investigate this relationship, photosynthetic parameters were measured in maize ( Zea mays , C4) and basil ( Ocimum basilicum , C3) under different photon densities and spectral compositions. Maize showed the highest CO 2 assimilation rate of 30.99 ± 1.54 µmol CO 2 /(m²s) under 2000 PAR green light (527 nm), while basil reached 10.56 ± 0.92 µmol CO 2 /(m²s) under red light (630 nm). PAM-derived electron transport rates (ETR) increased with light intensity in a pattern similar to CO 2 assimilation, but did not reliably reflect its absolute values under all conditions. To improve prediction accuracy, we applied a machine learning model. XGBoost, a gradient-boosted decision tree algorithm, efficiently captures nonlinear interactions between physiological and environmental parameters. It achieved superior performance (R² = 0.847; MSE = 5.24) compared to the Random Forest model. Our model enables accurate photosynthesis prediction from PAM data across light intensities and spectral conditions in both C3 and C4 plants.

Why it matches plant phenotyping methodsPAM蛍光データから光合成性能を推定する機械学習モデルを開発・比較評価しており、植物生理形質の取得・推定手法が研究の中心である。

abstractTo improve prediction accuracy, we applied a machine learning model.

Image analysis using smartphones: relationship between leaf color and fresh weight of lettuce under different nutritional treatments. →

Image analysis can be useful for assessing crop health and predicting yield. Instead of expensive equipment, smartphones are considered an accessible and low-cost alternative. The objectives of this study were to evaluate whether fresh weight in green and red lettuce could be predicted by leaf color (intensity of green color measured by RGB) under different fertilizer treatments using RGB imaging from two widely used smartphone models (Samsung Galaxy and Apple iPhone). The two smartphones showed similar longitudinal patterns of RGB data (the intensity and dark green proportion), but the absolute difference in the RGB data was significantly different. Therefore, the averaged results were used for the analyses. Color intensity and dark green proportion were associated with the fresh lettuce weight (p = 0.005, 0.003, 0.014 and p < 0.001, respectively). This study suggests that farmers and practitioners can use these economic devices as a non-destructive method to diagnose and monitor the nutritional status and predict lettuce yield.

Why it matches plant phenotyping methodsスマートフォンRGB画像から葉色を抽出し、レタスの生体重・栄養状態を非破壊推定する手法が研究の中心であり、植物表現型取得への実質的な応用に該当する。

abstractsmartphones are considered an accessible and low-cost alternative

Seed-to-plant-tracking: automated phenotyping of seeds and corresponding plants of Arabidopsis →

Plants adapt seed traits in response to different environmental triggers, supporting the survival of the next generation. To elucidate the mechanistic understanding of such adaptations it is important to characterize the distributions of seed traits by phenotyping seeds on an individual scale and to correlate these traits with corresponding plant properties. Here we introduce a seed-to-plant-tracking pipeline which enables automated handling and high precision phenotyping of Arabidopsis seeds as well as germination detection and early growth quantification of emerging plants. It includes previously published measurement platforms ( pheno Seeder, Growscreen), which were improved for very small seeds. We demonstrate the performance of the pipeline by comparing seeds from two consecutive generations of elevated temperature during flowering with control seeds. Relative standard deviation of repeated seed mass measurements was reduced to 0.2%. We identified an increase in seed mass, volume, length, width, height, and germination time as well as a darkening of the seeds under the treatment. A correlation analysis revealed relationships between seed and plant traits, e.g., a highly significant negative correlation between seed brightness and germination time, and a positive correlation between seed mass and early growth rate, but no correlation between time of emergence and morphometric seed traits (e.g., mass, volume). Thus, the seed-to-plant tracking provides the basis for investigating the mechanism of seed and plant trait variation and transgenerational inheritance.

Why it matches plant phenotyping methods種子から植物までを追跡し、種子形質の高精度自動計測、発芽検出、初期成長定量を行うパイプラインを開発・改良しており、表現型取得法が研究の中心である。

abstractHere we introduce a seed-to-plant-tracking pipeline which enables automated handling and high precision phenotyping of Arabidopsis seeds as well as germination detection and early growth quantification of emerging plants.

TomaFDNet: A multiscale focused diffusion-based model for tomato disease detection. →

Introduction Tomatoes are one of the most economically significant crops worldwide, with their yield and quality heavily impacted by foliar diseases. Effective detection of these diseases is essential for enhancing agricultural productivity and mitigating economic losses. Current tomato leaf disease detection methods, however, encounter challenges in extracting multi-scale features, identifying small targets, and mitigating complex background interference. Methods The multi-scale tomato leaf disease detection model Tomato Focus-Diffusion Network (TomaFDNet) was proposed to solve the above problems. The model utilizes a multi-scale focus-diffusion network (MSFDNet) alongside an efficient parallel multi-scale convolutional module (EPMSC) to significantly enhance the extraction of multi-scale features. This combination particularly strengthens the model's capability to detect small targets amidst complex backgrounds. Results and discussion Experimental results show that TomaFDNet reaches a mean average precision (mAP) of 83.1% in detecting Early_blight, Late_blight, and Leaf_Mold on tomato leaves, outperforming classical object detection algorithms, including Faster R-CNN (mAP = 68.2%) and You Only Look Once (YOLO) series (v5: mAP = 75.5%, v7: mAP = 78.3%, v8: mAP = 78.9%, v9: mAP = 79%, v10: mAP = 77.5%, v11: mAP = 79.2%). Compared to the baseline YOLOv8 model, TomaFDNet achieves a 4.2% improvement in mAP, which is statistically significant (P < 0.01). These findings indicate that TomaFDNet offers a valid solution to the precise detection of tomato leaf diseases.

Why it matches plant phenotyping methodsトマト葉の病徴を画像から検出する新規モデルを開発し、既存手法との性能比較で技術的に検証しているため、植物病害状態のフェノタイピング手法が中心である。

abstractThe multi-scale tomato leaf disease detection model Tomato Focus-Diffusion Network (TomaFDNet) was proposed to solve the above problems.

Identification of rice leaf disease based on DepMulti-Net. →

This research presents DepMulti-Net, a novel rice disease and pest identification model, designed to overcome the challenges of complex background interference, difficult disease feature extraction, and large model parameter volume in rice leaf disease identification. Initially, a comprehensive rice disease dataset comprising 20,000 images was meticulously constructed, covering four common types of rice diseases: bacterial leaf blight, rice blast, brown spot, and tungro disease. To enhance data diversity, various data augmentation techniques were applied. Subsequently, a novel VGG-block module was introduced. By leveraging depth-separable convolution, the model's parameter quantity was significantly reduced. A multi-scale feature fusion module was also designed to effectively enhance the model's ability to extract disease features from complex backgrounds. Moreover, the integration of the feature reuse mechanism and inverse bottleneck structure further improved the model's recognition accuracy for fine-grained disease features. Experimental results show that the DepMulti-Net model has only 13.50M parameters and achieves an average accuracy of 98.56% in identifying the four types of rice diseases. This performance significantly outperforms existing rice leaf disease identification methods. In conclusion, this study offers an efficient and lightweight solution for crop disease identification, which holds great significance for promoting the development of smart agriculture.

Why it matches plant phenotyping methodsイネ葉の病徴を画像から識別する深層学習モデルを開発し、データセット構築と性能評価を行っており、植物病害状態の表現型推定手法が中心である。

abstractThis research presents DepMulti-Net, a novel rice disease and pest identification model

StatFaRmer: cultivating insights with an advanced R shiny dashboard for digital phenotyping data analysis. →

Digital phenotyping is a fast-growing area of hardware and software research and development. Phenotypic studies usually require determining whether there is a difference in some trait between plants with different genotypes or under different conditions. We developed StatFaRmer, a user-friendly tool tailored for analyzing time series of plant phenotypic parameters, ensuring seamless integration with common tasks in phenotypic studies. For maximum versatility across phenotypic methods and platforms, it uses data in the form of a set of spreadsheets (XLSX and CSV files). StatFaRmer is designed to handle measurements that have variation in timestamps between plants and the presence of outliers, which is common in digital phenotyping. Data preparation is automated and well-documented, leading to customizable ANOVA tests that include diagnostics and significance estimation for effects between user-defined groups. Users can download the results from each stage and reproduce their analysis. It was tested and shown to work reliably for large datasets across various experimental designs with a wide range of plants, including bread wheat (Triticum aestivum), durum wheat (Triticum durum), and triticale (× Triticosecale); sugar beet (Beta vulgaris), cocklebur (Xanthium strumarium) and lettuce (Lactuca sativa), corn (Zea mays) and sunflower (Helianthus annuus), and soybean (Glycine max). StatFaRmer is created as an open-source Shiny dashboard, and simple instructions on installation and operation on Windows and Linux are provided.

Why it matches plant phenotyping methods植物フェノタイピングの時系列データ解析を目的とするオープンソースShinyダッシュボードを開発し、データ準備・統計解析・再現可能なワークフローを提供しており、方法・ソフトウェアが中心である。

abstractWe developed StatFaRmer, a user-friendly tool tailored for analyzing time series of plant phenotypic parameters

Image-based yield prediction for tall fescue using random forests and convolutional neural networks →

In the early stages of selection, many plant breeding programmes still rely on visual evaluations of traits by experienced breeders. While this approach has proven to be effective, it requires considerable time, labour and expertise. Moreover, its subjective nature makes it difficult to reproduce and compare evaluations. The field of automated high-throughput phenotyping aims to resolve these issues. A widely adopted strategy uses drone images processed by machine learning algorithms to characterise phenotypes. This approach was used in the present study to assess the dry matter yield of tall fescue and its accuracy was compared to that of the breeder's evaluations, using field measurements as ground truth. RGB images of tall fescue individuals were processed by two types of predictive models: a random forest and convolutional neural network. In addition to computing dry matter yield, the two methods were applied to identify the top 10% highest-yielding plants and predict the breeder's score. The convolutional neural network outperformed the random forest method and exceeded the predictive power of the breeder's eye. It predicted dry matter yield with an R² of 0.62, which surpassed the accuracy of the breeder's score by 8 percentage points. Additionally, the algorithm demonstrated strong performance in identifying top-performing plants and estimating the breeder's score, achieving balanced accuracies of 0.81 and 0.74, respectively. These findings indicate that the tested automated phenotyping approach could not only offer improvements in cost, time efficiency and objectivity, but also enhance selection accuracy. As a result, this technique has the potential to increase overall breeding efficiency, accelerate genetic progress, and shorten the time to market. To conclude, phenotyping by means of RGB-based machine learning models provides a reliable alternative or addition to the visual evaluation of selection candidates in a tall fescue breeding programme.

Why it matches plant phenotyping methodsRGBドローン画像と機械学習により乾物収量などの植物形質を推定し、育種家評価および実測値と比較検証しており、フェノタイピング手法が中心である。

abstractThe field of automated high-throughput phenotyping aims to resolve these issues.

DM-YOLO: improved YOLOv9 model for tomato leaf disease detection. →

In natural environments, tomato leaf disease detection faces many challenges, such as variations in light conditions, overlapping disease symptoms, tiny size of lesion areas, and occlusion between leaves. Therefore, an improved tomato leaf disease detection method, DM-YOLO, based on the YOLOv9 algorithm, is proposed in this paper. Specifically, firstly, lightweight dynamic up-sampling DySample is incorporated into the feature fusion backbone network to enhance the ability to extract features of small lesions and suppress the interference from the background environment; secondly, the MPDIoU loss function is used to enhance the learning of the details of overlapping lesion margins in order to improve the accuracy of localizing overlapping lesion margins. The experimental results show that the precision (P) of this model increased by 2.2%, 1.7%, 2.3%, 2%, and 2.1%compared with those of multiple mainstream improved models, respectively. When evaluated based on the tomato leaf disease dataset, the precision (P) of the model was 92.5%, and the average precision (AP) and the mean average precision (mAP) were 95.1% and 86.4%, respectively, which were 3%, 1.7%, and 1.4% higher than the P, AP, and mAP of YOLOv9, the baseline model, respectively. The proposed detection method had good detection performance and detection potential, which will provide strong support for the development of smart agriculture and disease control.

Why it matches plant phenotyping methodsトマト葉の病斑を画像から検出するYOLO改良手法の開発と性能評価が中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractan improved tomato leaf disease detection method, DM-YOLO, based on the YOLOv9 algorithm, is proposed in this paper.

Tea disease identification based on ECA attention mechanism ResNet50 network. →

Addressing the challenge of identifying tea plant diseases against the complex background of tea gardens, this study proposes the ECA-ResNet50 model. By optimizing the ResNet50 architecture, adopting a multi-layer small convolution kernel strategy to enhance feature extraction capabilities, and introducing the ECA attention mechanism to focus on key features, the model achieves a 93.06% accuracy rate in tea disease identification, representing a 3.18% improvement over the original model, demonstrating industry-leading performance advantages. This model not only accurately identifies tea diseases in gardens but also possesses excellent generalization capabilities, performing outstandingly on datasets of other plant categories. These results indicate that ECA-ResNet50 can effectively mitigate the interference of complex backgrounds and precisely recognize tea disease targets.

Why it matches plant phenotyping methods茶葉の病害を画像から識別する深層学習モデルを開発し、精度比較と汎化性能を評価しており、植物病害状態の取得・推定手法が中心である。

abstractthis study proposes the ECA-ResNet50 model.

Adaptive spatial-channel feature fusion and self-calibrated convolution for early maize seedlings counting in UAV images. →

Accurate counting of crop plants is essential for agricultural science, particularly for yield forecasting, field management, and experimental studies. Traditional methods are labor-intensive and prone to errors. Unmanned Aerial Vehicle (UAV) technology offers a promising alternative; however, varying UAV altitudes can impact image quality, leading to blurred features and reduced accuracy in early maize seedling counts. To address these challenges, we developed RC-Dino, a deep learning methodology based on DINO, specifically designed to enhance the precision of seedling counts from UAV-acquired images. RC-Dino introduces two innovative components: a novel self-calibrating convolutional layer named RSCconv and an adaptive spatial feature fusion module called ASCFF. The RSCconv layer improves the representation of early maize seedlings compared to non-seedling elements within feature maps by calibrating spatial domain features. The ASCFF module enhances the discriminability of early maize seedlings by adaptively fusing feature maps extracted from different layers of the backbone network. Additionally, transfer learning was employed to integrate pre-trained weights with RSCconv, facilitating faster convergence and improved accuracy. The efficacy of our approach was validated using the Early Maize Seedlings Dataset (EMSD), comprising 1,233 annotated images of early maize seedlings, totaling 83,404 individual annotations. Testing on this dataset demonstrated that RC-Dino outperformed existing models, including DINO, Faster R-CNN, RetinaNet, YOLOX, and Deformable DETR. Specifically, RC-Dino achieved improvements of 16.29% in Average Precision (AP) and 8.19% in Recall compared to the DINO model. Our method also exhibited superior coefficient of determination (R²) values across different datasets for seedling counting. By integrating RSCconv and ASCFF into other detection frameworks such as Faster R-CNN, RetinaNet, and Deformable DETR, we observed enhanced detection and counting accuracy, further validating the effectiveness of our proposed method. These advancements make RC-Dino particularly suitable for accurate early maize seedling counting in the field. The source code for RSCconv and ASCFF is publicly available at https://github.com/collapser-AI/RC-Dino, promoting further research and practical applications.

Why it matches plant phenotyping methodsUAV画像からトウモロコシ幼苗数を抽出する深層学習手法を開発し、公開データセット上で既存手法と比較検証しており、植物表現型取得法が研究の中心です。

abstractwe developed RC-Dino, a deep learning methodology based on DINO, specifically designed to enhance the precision of seedling counts from UAV-acquired images.

An improved ShuffleNetV2 method based on ensemble self-distillation for tomato leaf diseases recognition. →

Introduction Timely and accurate recognition of tomato diseases is crucial for improving tomato yield. While large deep learning models can achieve high-precision disease recognition, these models often have a large number of parameters, making them difficult to deploy on edge devices. To address this issue, this study proposes an ensemble self-distillation method and applies it to the lightweight model ShuffleNetV2. Methods Specifically, based on the architecture of ShuffleNetV2, multiple shallow models at different depths are constructed to establish a distillation framework. Based on the fused feature map that integrates the intermediate feature maps of ShuffleNetV2 and shallow models, a depthwise separable convolution layer is introduced to further extract more effective feature information. This method ensures that the intermediate features from each model are fully preserved to the ensemble model, thereby improving the overall performance of the ensemble model. The ensemble model, acting as the teacher, dynamically transfers knowledge to ShuffleNetV2 and the shallow models during training, significantly enhancing the performance of ShuffleNetV2 without changing the original structure. Results Experimental results show that the optimized ShuffleNetV2 achieves an accuracy of 95.08%, precision of 94.58%, recall of 94.55%, and an F1 score of 94.54% on the test set, surpassing large models such as VGG16 and ResNet18. Among lightweight models, it has the smallest parameter count and the highest recognition accuracy. Discussion The results demonstrate that the optimized ShuffleNetV2 is more suitable for deployment on edge devices for real-time tomato disease detection. Additionally, multiple shallow models achieve varying degrees of compression for ShuffleNetV2, providing flexibility for model deployment.

Why it matches plant phenotyping methodsトマト葉の病徴を画像から認識する深層学習モデルを開発・評価しており、植物病害状態の取得・推定手法が研究の中心である。

abstractthis study proposes an ensemble self-distillation method and applies it to the lightweight model ShuffleNetV2.

Classification of tomato leaf disease using Transductive Long Short-Term Memory with an attention mechanism. →

Tomatoes are considered one of the most valuable vegetables around the world due to their usage and minimal harvesting period. However, effective harvesting still remains a major issue because tomatoes are easily susceptible to weather conditions and other types of attacks. Thus, numerous research studies have been introduced based on deep learning models for the efficient classification of tomato leaf disease. However, the usage of a single architecture does not provide the best results due to the limited computational ability and classification complexity. Thus, this research used Transductive Long Short-Term Memory (T-LSTM) with an attention mechanism. The attention mechanism introduced in T-LSTM has the ability to focus on various parts of the image sequence. Transductive learning exploits the specific characteristics of the training instances to make accurate predictions. This can involve leveraging the relationships and patterns observed within the dataset. The T-LSTM is based on the transductive learning approach and the scaled dot product attention evaluates the weights of each step based on the hidden state and image patches which helps in effective classification. The data was gathered from the PlantVillage dataset and the pre-processing was conducted based on image resizing, color enhancement, and data augmentation. These outputs were then processed in the segmentation stage where the U-Net architecture was applied. After segmentation, VGG-16 architecture was used for feature extraction and the classification was done through the proposed T-LSTM with an attention mechanism. The experimental outcome shows that the proposed classifier achieved an accuracy of 99.98% which is comparably better than existing convolutional neural network models with transfer learning and IBSA-NET.

Why it matches plant phenotyping methodsトマト葉の病害状態を画像から推定する分類・セグメンテーション手法を提案し、前処理、U-Net、特徴抽出、T-LSTM分類、比較評価までが中心であるため、植物病害フェノタイピング手法として含める。

abstractThese outputs were then processed in the segmentation stage where the U-Net architecture was applied.

DFMA: an improved DeepLabv3+ based on FasterNet, multi-receptive field, and attention mechanism for high-throughput phenotyping of seedlings. →

With the rapid advancement of plant phenotyping research, understanding plant genetic information and growth trends has become crucial. Measuring seedling length is a key criterion for assessing seed viability, but traditional ruler-based methods are time-consuming and labor-intensive. To address these limitations, we propose an efficient deep learning approach to enhance plant seedling phenotyping analysis. We improved the DeepLabv3+ model, naming it DFMA, and introduced a novel ASPP structure, PSPA-ASPP. On our self-constructed rice seedling dataset, the model achieved a mean Intersection over Union (mIoU) of 81.72%. On publicly available datasets, including Arabidopsis thaliana, Brachypodium distachyon, and Sinapis alba, detection scores reached 87.69%, 91.07%, and 66.44%, respectively, outperforming existing models. The model generates detailed segmentation masks, capturing structures such as the embryonic shoot, axis, and root, while a seedling length measurement algorithm provides precise parameters for component development. This approach offers a comprehensive, automated solution, improving phenotyping analysis efficiency and addressing the challenges of traditional methods.

Why it matches plant phenotyping methods幼苗画像から器官を分割し、幼苗長を自動測定する深層学習ベースのフェノタイピング手法を開発・検証しており、方法が中心的です。

abstractwe propose an efficient deep learning approach to enhance plant seedling phenotyping analysis.

Advanced deep transfer learning techniques for efficient detection of cotton plant diseases →

Introduction Cotton, being a crucial cash crop globally, faces significant challenges due to multiple diseases that adversely affect its quality and yield. To identify such diseases is very important for the implementation of effective management strategies for sustainable agriculture. Image recognition plays an important role for the timely and accurate identification of diseases in cotton plants as it allows farmers to implement effective interventions and optimize resource allocation. Additionally, deep learning has begun as a powerful technique for to detect diseases in crops using images. Hence, the significance of this work lies in its potential to mitigate the impact of these diseases, which cause significant damage to the cotton and decrease fibre quality and promote sustainable agricultural practices. Methods This paper investigates the role of deep transfer learning techniques such as EfficientNet models, Xception, ResNet models, Inception, VGG, DenseNet, MobileNet, and InceptionResNet for cotton plant disease detection. A complete dataset of infected cotton plants having diseases like Bacterial Blight, Target Spot, Powdery Mildew, Aphids, and Army Worm along with the healthy ones is used. After pre-processing the images of the dataset, their region of interest is obtained by applying feature extraction techniques such as the generation of the biggest contour, identification of extreme points, cropping of relevant regions, and segmenting the objects using adaptive thresholding. Results and Discussion During experimentation, it is found that the EfficientNetB3 model outperforms in accuracy, loss, as well as root mean square error by obtaining 99.96%, 0.149, and 0.386 respectively. However, other models also show the good performance in terms of precision, recall, and F1 score, with high scores close to 0.98 or 1.00, except for VGG19. The findings of the paper emphasize the prospective of deep transfer learning as a viable technique for cotton plant disease diagnosis by providing a cost-effective and efficient solution for crop disease monitoring and management. This strategy can also help to improve agricultural practices by ensuring sustainable cotton farming and increased crop output.

Why it matches plant phenotyping methods綿花の画像から植物病害状態を推定する深層学習手法を比較・評価しており、病害フェノタイピング手法が中心である。

abstractThis paper investigates the role of deep transfer learning techniques such as EfficientNet models, Xception, ResNet models, Inception, VGG, DenseNet, MobileNet, and InceptionResNet for cotton plant disease detection.

Disentangling genotype and environment specific latent features for improved trait prediction using a compositional autoencoder. →

In plant breeding and genetics, predictive models traditionally rely on compact representations of high-dimensional data, often using methods like Principal Component Analysis (PCA) and, more recently, Autoencoders (AE). However, these methods do not separate genotype-specific and environment-specific features, limiting their ability to accurately predict traits influenced by both genetic and environmental factors. We hypothesize that disentangling these representations into genotype-specific and environment-specific components can enhance predictive models. To test this, we developed a compositional autoencoder (CAE) that decomposes high-dimensional data into distinct genotype-specific and environment-specific latent features. Our CAE framework employed a hierarchical architecture within an autoencoder to effectively separate these entangled latent features. Applied to a maize diversity panel dataset, the CAE demonstrated superior modeling of environmental influences and out-performs PCA (principal component analysis), PLSR (Partial Least square regression) and vanilla autoencoders by 7 times for 'Days to Pollen' trait and 10 times improved predictive performance for 'Yield'. By disentangling latent features, the CAE provided a powerful tool for precision breeding and genetic research. This work has significantly enhanced trait prediction models, advancing agricultural and biological sciences.

Why it matches plant phenotyping methods植物形質（開花日数・収量）の予測を目的とする新規オートエンコーダを開発し、既存手法と比較評価しており、計算的な形質推定手法が研究の中心である。

abstractwe developed a compositional autoencoder (CAE) that decomposes high-dimensional data into distinct genotype-specific and environment-specific latent features.

DINOV2-FCS: a model for fruit leaf disease classification and severity prediction. →

Introduction The assessment of the severity of fruit disease is crucial for the optimization of fruit production. By quantifying the percentage of leaf disease, an effective approach to determining the severity of the disease is available. However, the current prediction of disease degree by machine learning methods still faces challenges, including suboptimal accuracy and limited generalizability. Methods In light of the growing application of large model technology across a range of fields, this study draws upon the DINOV2 visual large vision model backbone network to construct the DINOV2-Fruit Leaf Classification and Segmentation Model (DINOV2-FCS), a model designed for the classification and severity prediction of diverse fruit leaf diseases. DINOV2-FCS employs the DINOv2-B (distilled) backbone feature extraction network to enhance the extraction of features from fruit disease leaf images. In fruit leaf disease classification, for the problem that leaf spots of different diseases have great similarity, we have proposed Class-Patch Feature Fusion Module (C-PFFM), which integrates the local detailed feature information of the spots and the global feature information of the class markers. For the problem that the model ignores the fine spots in the segmentation process, we propose Explicit Feature Fusion Architecture (EFFA) and Alterable Kernel Atrous Spatial Pyramid Pooling (AKASPP), which improve the segmentation effect of the model. Results To verify the accuracy and generalizability of the model, two sets of experiments were conducted. First, the labeled leaf disease dataset of five fruits was randomly divided. The trained model exhibited an accuracy of 99.67% in disease classification, an mIoU of 90.29%, and an accuracy of 95.68% in disease severity classification. In the generalizability experiment, four disease data sets were used for training and one for testing. The mIoU of the trained model reached 83.95%, and the accuracy of disease severity grading was 95.24%. Discussion The results demonstrate that the model exhibits superior performance compared to other state-of-the-art models and that the model has strong generalization capabilities. This study provides a new method for leaf disease classification and leaf disease severity prediction for a variety of fruits. Code is available at https://github.com/BaiChunhui2001/DINOV2-FCS.

Why it matches plant phenotyping methods果樹葉の病害分類と病害重症度を画像から推定する深層学習モデルを開発し、精度・汎化性能を検証しており、植物病態表現型の取得・推定手法が中心である。

abstracta model designed for the classification and severity prediction of diverse fruit leaf diseases

AppleLeafNet: a lightweight and efficient deep learning framework for diagnosing apple leaf diseases. →

Accurately identifying apple diseases is essential to control their spread and support the industry. Timely and precise detection is crucial for managing the spread of diseases, thereby improving the production and quality of apples. However, the development of algorithms for analyzing complex leaf images remains a significant challenge. Therefore, in this study, a lightweight deep learning model is designed from scratch to identify the apple leaf condition. The developed framework comprises two stages. First, the designed 37-layer model was employed to assess the condition of apple leaves (healthy or diseased). Second, transfer learning was used for further subclassification of the disease class (e.g., rust, complex, scab, and frogeye leaf spots). The trained lightweight model was reused because the model trained with correlated images facilitated transfer learning for further classification of the disease class. A dataset available online was used to validate the proposed two-stage framework, resulting in a classification rate of 98.25% for apple leaf condition identification and an accuracy of 98.60% for apple leaf disease diagnosis. Furthermore, the results confirm that the proposed model is lightweight and involves relatively fewer learnable parameters in comparison with other pre-trained deep learning models.

Why it matches plant phenotyping methodsリンゴ葉画像から健全・罹病状態および病害クラスを推定する軽量深層学習手法を開発し、精度検証しており、植物病害表現型の取得・分類が中心である。

abstracta lightweight deep learning model is designed from scratch to identify the apple leaf condition

MS-YOLOv8: multi-scale adaptive recognition and counting model for peanut seedlings under salt-alkali stress from remote sensing. →

Introduction The emergence rate of crop seedlings is an important indicator for variety selection, evaluation, field management, and yield prediction. To address the low recognition accuracy caused by the uneven size and varying growth conditions of crop seedlings under salt-alkali stress, this research proposes a peanut seedling recognition model, MS-YOLOv8. Methods This research employs close-range remote sensing from unmanned aerial vehicles (UAVs) to rapidly recognize and count peanut seedlings. First, a lightweight adaptive feature fusion module (called MSModule) is constructed, which groups the channels of input feature maps and feeds them into different convolutional layers for multi-scale feature extraction. Additionally, the module automatically adjusts the channel weights of each group based on their contribution, improving the feature fusion effect. Second, the neck network structure is reconstructed to enhance recognition capabilities for small objects, and the MPDIoU loss function is introduced to effectively optimize the detection boxes for seedlings with scattered branch growth. Results Experimental results demonstrate that the proposed MS-YOLOv8 model achieves an AP50 of 97.5% for peanut seedling detection, which is 12.9%, 9.8%, 4.7%, 5.0%, 11.2%, 5.0%, and 3.6% higher than Faster R-CNN, EfficientDet, YOLOv5, YOLOv6, YOLOv7, YOLOv8, and RT-DETR, respectively. Discussion This research provides valuable insights for crop recognition under extreme environmental stress and lays a theoretical foundation for the development of intelligent production equipment.

Why it matches plant phenotyping methodsUAVリモートセンシング画像からピーナッツ幼苗を認識・計数するモデルを開発し、検出性能を比較検証している。幼苗数・出現率という植物状態の推定が研究の中心である。

abstractthis research proposes a peanut seedling recognition model, MS-YOLOv8

ANFIS Fuzzy convolutional neural network model for leaf disease detection →

Leaf disease detection is critical in agriculture, as it directly impacts crop health, yield, and quality. Early and accurate detection of leaf diseases can prevent the spread of infections, reduce the need for chemical treatments, and minimize crop losses. This not only ensures food security but also supports sustainable farming practices. Effective leaf disease detection systems empower farmers with the knowledge to take timely actions, leading to healthier crops and more efficient resource management. In an era of increasing global food demand and environmental challenges, advanced leaf disease detection technologies are indispensable for modern agriculture. This study presents an innovative approach for detecting pepper bell leaf disease using an ANFIS Fuzzy convolutional neural network (CNN) integrated with local binary pattern (LBP) features. Experiments involve using the models without LBP, as well as, with LBP features. For both sets of experiments, the proposed ANFIS CNN model performs superbly. It shows an accuracy score of 0.8478 without using LBP features while its precision, recall, and F1 scores are 0.8959, 0.9045, and 0.8953, respectively. Incorporating LBP features, the proposed model achieved exceptional performance, with accuracy, precision, recall, and an F1 score of higher than 99%. Comprehensive comparisons with state-of-the-art techniques further highlight the superiority of the proposed method. Additionally, cross-validation was applied to ensure the robustness and reliability of the results. This approach demonstrates a significant advancement in agricultural disease detection, promising enhanced accuracy and efficiency in real-world applications.

Why it matches plant phenotyping methods画像からコショウ葉の病害状態を推定するCNN手法を開発・比較・検証しており、植物病害フェノタイピング手法が中心です。

abstractThis study presents an innovative approach for detecting pepper bell leaf disease using an ANFIS Fuzzy convolutional neural network (CNN) integrated with local binary pattern (LBP) features.

LCGSC-YOLO: a lightweight apple leaf diseases detection method based on LCNet and GSConv module under YOLO framework. →

Introduction In response to the current mainstream deep learning detection methods with a large number of learned parameters and the complexity of apple leaf disease scenarios, the paper proposes a lightweight method and names it LCGSC-YOLO. This method is based on the LCNet(A Lightweight CPU Convolutional Neural Network) and GSConv(Group Shuffle Convolution) module modified YOLO(You Only Look Once) framework. Methods Firstly, the lightweight LCNet is utilized to reconstruct the backbone network, with the purpose of reducing the number of parameters and computations of the model. Secondly, the GSConv module and the VOVGSCSP (Slim-neck by GSConv) module are introduced in the neck network, which makes it possible to minimize the number of model parameters and computations while guaranteeing the fusion capability among the different feature layers. Finally, coordinate attention is embedded in the tail of the backbone and after each VOVGSCSP module to improve the problem of detection accuracy degradation issue caused by model lightweighting. Results The experimental results show the LCGSC-YOLO can achieve an excellent detection performance with mean average precision of 95.5% and detection speed of 53 frames per second (FPS) on the mixed datasets of Plant Pathology 2021 (FGVC8) and AppleLeaf9. Discussion The number of parameters and Floating Point Operations (FLOPs) of the LCGSC-YOLO are much less thanother related comparative experimental algorithms.

Why it matches plant phenotyping methodsリンゴ葉の病害症状を画像から検出する軽量な深層学習手法を開発し、精度と処理速度を評価しており、植物表現型取得が中心的貢献である。

abstractthe paper proposes a lightweight method and names it LCGSC-YOLO.

Of buds and bits: a meta-QTL study identifies stable QTL for berry quality and yield traits in cranberry mapping populations ( Vaccinium macrocarpon Ait.). →

Introduction For nearly two centuries, cranberry (Vaccinium macrocarpon Ait.) breeders have improved fruit quality and yield by selecting traits on fruiting stems, termed “reproductive uprights.” Crop improvement is accelerating rapidly in contemporary breeding programs due to modern genetic tools and high-throughput phenotyping methods, improving selection efficiency and accuracy. Methods We conducted genotypic evaluation on 29 primary traits encompassing fruit quality, yield, and chemical composition in two full-sib cranberry breeding populations—CNJ02 (n = 168) and CNJ04 (n = 67)—over 3 years. Genetic characterization was further performed on 11 secondary traits derived from these primary traits. Results For CNJ02, 170 major quantitative trait loci (QTL; R2≥ 0.10) were found with interval mapping, 150 major QTL were found with model mapping, and 9 QTL were found to be stable across multiple years. In CNJ04, 69 major QTL were found with interval mapping, 81 major QTL were found with model mapping, and 4 QTL were found to be stable across multiple years. Meta-QTL represent stable genomic regions consistent across multiple years, populations, studies, or traits. Seven multi-trait meta-QTL were found in CNJ02, one in CNJ04, and one in the combined analysis of both populations. A total of 22 meta-QTL were identified in cross-study, cross-population analysis using digital traits for berry shape and size (8 meta-QTL), digital images for berry color (2 meta-QTL), and three-study cross-analysis (12 meta-QTL). Discussion Together, these meta-QTL anchor high-throughput fruit quality phenotyping techniques to traditional phenotyping methods, validating state-of-the-art methods in cranberry phenotyping that will improve breeding accuracy, efficiency, and genetic gain in this globally significant fruit crop.

Why it matches plant phenotyping methodsベリー形状・サイズ・色のデジタル形質および画像を用いた果実フェノタイピングをQTL解析で検証し、従来法との対応付けを行っており、手法の妥当性評価が研究の主要な成果に含まれる。

abstractdigital traits for berry shape and size (8 meta-QTL), digital images for berry color (2 meta-QTL)

An optimized live imaging and multiple cell layer growth analysis approach using Arabidopsis sepals. →

Arabidopsis thaliana sepals are excellent models for analyzing growth of entire organs due to their relatively small size, which can be captured at a cellular resolution under a confocal microscope. To investigate how differential growth of connected cell layers generate unique organ morphologies, it is necessary to live-image deep into the tissue. However, imaging deep cell layers of the sepal (or plant tissues in general) is practically challenging. Image processing is also difficult due to the low signal-to-noise ratio of the deeper tissue layers, an issue mainly associated with live imaging datasets. Addressing some of these challenges, we provide an optimized methodology for live imaging sepals, and subsequent image processing. For live imaging early-stage sepals, we found that the use of a bright fluorescent membrane marker, coupled with increased laser intensity and an enhanced Z- resolution produces high-quality images suitable for downstream image processing. Our optimized parameters allowed us to image the bottommost cell layer of the sepal (inner epidermal layer) without compromising viability. We used a 'voxel removal' technique to visualize the inner epidermal layer in MorphoGraphX image processing software. We also describe the MorphoGraphX parameters for creating a 2.5D mesh surface for the inner epidermis. Our parameters allow for the segmentation and parent tracking of individual cells through multiple time points, despite the weak signal of the inner epidermal cells. While we have used sepals to illustrate our approach, the methodology will be useful for researchers intending to live-image and track growth of deeper cell layers in 2.5D for any plant tissue.

Why it matches plant phenotyping methods植物組織の深部をライブイメージングし、画像処理・細胞セグメンテーション・追跡によって成長を解析する方法自体が中心的に開発・最適化されているため。

abstractwe provide an optimized methodology for live imaging sepals, and subsequent image processing.

A lightweight dual-attention network for tomato leaf disease identification. →

Tomato disease image recognition plays a crucial role in agricultural production. Today, while machine vision methods based on deep learning have achieved some success in disease recognition, they still face several challenges. These include issues such as imbalanced datasets, unclear disease features, small inter-class differences, and large intra-class variations. To address these challenges, this paper proposes a method for classifying and recognizing tomato leaf diseases based on machine vision. First, to enhance the disease feature details in images, a piecewise linear transformation method is used for image enhancement, and oversampling is employed to expand the dataset, compensating for the imbalanced dataset. Next, this paper introduces a convolutional block with a dual attention mechanism called DAC Block, which is used to construct a lightweight model named LDAMNet. The DAC Block innovatively uses Hybrid Channel Attention (HCA) and Coordinate Attention (CSA) to process channel information and spatial information of input images respectively, enhancing the model's feature extraction capabilities. Additionally, this paper proposes a Robust Cross-Entropy (RCE) loss function that is robust to noisy labels, aimed at reducing the impact of noisy labels on the LDAMNet model during training. Experimental results show that this method achieves an average recognition accuracy of 98.71% on the tomato disease dataset, effectively retaining disease information in images and capturing disease areas. Furthermore, the method also demonstrates strong recognition capabilities on rice crop disease datasets, indicating good generalization performance and the ability to function effectively in disease recognition across different crops. The research findings of this paper provide new ideas and methods for the field of crop disease recognition. However, future research needs to further optimize the model's structure and computational efficiency, and validate its application effects in more practical scenarios.

Why it matches plant phenotyping methodsトマト葉の病害状態を画像から認識・領域抽出する深層学習手法を開発し、複数データセットで性能評価しており、植物フェノタイピング手法が中心である。

abstractthis paper proposes a method for classifying and recognizing tomato leaf diseases based on machine vision.

CT image-based 3D inflorescence estimation of Chrysanthemum seticuspe →

To study plant organs, it is necessary to investigate the three-dimensional (3D) structures of plants. In recent years, non-destructive measurements through computed tomography (CT) have been used to understand the 3D structures of plants. In this study, we use the Chrysanthemum seticuspe capitulum inflorescence as an example and focus on contact points between the receptacles and florets within the 3D capitulum inflorescence bud structure to investigate the 3D arrangement of the florets on the receptacle. To determine the 3D order of the contact points, we constructed slice images from the CT volume data and detected the receptacles and florets in the image. However, because each CT sample comprises hundreds of slice images to be processed and each C. seticuspe capitulum inflorescence comprises several florets, manually detecting the receptacles and florets is labor-intensive. Therefore, we propose an automatic contact point detection method based on CT slice images using image recognition techniques. The proposed method improves the accuracy of contact point detection using prior knowledge that contact points exist only around the receptacle. In addition, the integration of the detection results enables the estimation of the 3D position of the contact points. According to the experimental results, we confirmed that the proposed method can detect contacts on slice images with high accuracy and estimate their 3D positions through clustering. Additionally, the sample-independent experiments showed that the proposed method achieved the same detection accuracy as sample-dependent experiments.

Why it matches plant phenotyping methodsCT画像から花序内の小花と花托の接触点を自動検出し、3D位置を推定する手法の開発・精度評価が研究の中心であるため、植物フェノタイピング手法に該当する。

abstractTherefore, we propose an automatic contact point detection method based on CT slice images using image recognition techniques.

Petal segmentation in CT images based on divide-and-conquer strategy. →

Manual segmentation of the petals of flower computed tomography (CT) images is time-consuming and labor-intensive because the flower has many petals. In this study, we aim to obtain a three-dimensional (3D) structure of Camellia japonica flowers and propose a petal segmentation method using computer vision techniques. Petal segmentation on the slice images fails by simply applying the segmentation methods because the shape of the petals in CT images differs from that of the objects targeted by the latest instance segmentation methods. To overcome these challenges, we crop two-dimensional (2D) long rectangles from each slice image and apply the segmentation method to segment the petals on the images. Thanks to cropping, it is easier to segment the shape of the petals in the cropped images using the segmentation methods. We can also use the latest segmentation method for the task because the number of images used for training is augmented by cropping. Subsequently, the results are integrated into 3D to obtain 3D segmentation volume data. The experimental results show that the proposed method can segment petals on slice images with higher accuracy than the method without cropping. The 3D segmentation results were also obtained and visualized successfully.

Why it matches plant phenotyping methods花弁のCT画像から3D構造を抽出する画像セグメンテーション手法の開発と精度比較が中心であり、植物形態フェノタイピングに該当する。

abstractThe experimental results show that the proposed method can segment petals on slice images with higher accuracy than the method without cropping.

Development of a deep-learning phenotyping tool for analyzing image-based strawberry phenotypes. →

Introduction In strawberry farming, phenotypic traits (such as crown diameter, petiole length, plant height, flower, leaf, and fruit size) measurement is essential as it serves as a decision-making tool for plant monitoring and management. To date, strawberry plant phenotyping has relied on traditional approaches. In this study, an image-based Strawberry Phenotyping Tool (SPT) was developed using two deep-learning (DL) architectures, namely “YOLOv4” and “U-net” integrated into a single system. We aimed to create the most suitable DL-based tool with enhanced robustness to facilitate digital strawberry plant phenotyping directly at the natural scene or indirectly using captured and stored images. Methods Our SPT was developed primarily through two steps (subsequently called versions) using image data with different backgrounds captured with simple smartphone cameras. The two versions (V1 and V2) were developed using the same DL networks but differed by the amount of image data and annotation method used during their development. For V1, 7,116 images were annotated using the single-target non-labeling method, whereas for V2, 7,850 images were annotated using the multitarget labeling method. Results The results of the held-out dataset revealed that the developed SPT facilitates strawberry phenotype measurements. By increasing the dataset size combined with multitarget labeling annotation, the detection accuracy of our system changed from 60.24% in V1 to 82.28% in V2. During the validation process, the system was evaluated using 70 images per phenotype and their corresponding actual values. The correlation coefficients and detection frequencies were higher for V2 than for V1, confirming the superiority of V2. Furthermore, an image-based regression model was developed to predict the fresh weight of strawberries based on the fruit size (R2 = 0.92). Discussion The results demonstrate the efficiency of our system in recognizing the aforementioned six strawberry phenotypic traits regardless of the complex scenario of the environment of the strawberry plant. This tool could help farmers and researchers make accurate and efficient decisions related to strawberry plant management, possibly causing increased productivity and yield potential.

Why it matches plant phenotyping methods画像ベースでイチゴの複数形質を抽出・測定する深層学習ツールを開発し、精度と実測値との相関を検証しており、植物表現型取得手法が研究の中心である。

abstractIn this study, an image-based Strawberry Phenotyping Tool (SPT) was developed using two deep-learning (DL) architectures, namely “YOLOv4” and “U-net” integrated into a single system.

Lightweight tomato ripeness detection algorithm based on the improved RT-DETR. →

Tomatoes, widely cherished for their high nutritional value, necessitate precise ripeness identification and selective harvesting of mature fruits to significantly enhance the efficiency and economic benefits of tomato harvesting management. Previous studies on intelligent harvesting often focused solely on identifying tomatoes as the target, lacking fine-grained detection of tomato ripeness. This deficiency leads to the inadvertent harvesting of immature and rotten fruits, resulting in economic losses. Moreover, in natural settings, uneven illumination, occlusion by leaves, and fruit overlap hinder the precise assessment of tomato ripeness by robotic systems. Simultaneously, the demand for high accuracy and rapid response in tomato ripeness detection is compounded by the need for making the model lightweight to mitigate hardware costs. This study proposes a lightweight model named PDSI-RTDETR to address these challenges. Initially, the PConv_Block module, integrating partial convolution with residual blocks, replaces the Basic_Block structure in the legacy backbone to alleviate computing load and enhance feature extraction efficiency. Subsequently, a deformable attention module is amalgamated with intra-scale feature interaction structure, bolstering the capability to extract detailed features for fine-grained classification. Additionally, the proposed slimneck-SSFF feature fusion structure, merging the Scale Sequence Feature Fusion framework with a slim-neck design utilizing GSConv and VoVGSCSP modules, aims to reduce volume of computation and inference latency. Lastly, by amalgamating Inner-IoU with EIoU to formulate Inner-EIoU, replacing the original GIoU to expedite convergence while utilizing auxiliary frames enhances small object detection capabilities. Comprehensive assessments validate that the PDSI-RTDETR model achieves an average precision mAP50 of 86.8%, marking a 3.9% enhancement over the original RT-DETR model, and a 38.7% increase in FPS. Furthermore, the GFLOPs of PDSI-RTDETR have been diminished by 17.6%. Surpassing the baseline RT-DETR and other prevalent methods regarding precision and speed, it unveils its considerable potential for detecting tomato ripeness. When applied to intelligent harvesting robots in the future, this approach can improve the quality of tomato harvesting by reducing the collection of immature and spoiled fruits.

Why it matches plant phenotyping methodsトマト果実の成熟度という植物状態を画像から推定する軽量検出モデルを開発し、精度・速度・計算量を評価しており、表現型取得手法が中心である。

abstractThis study proposes a lightweight model named PDSI-RTDETR to address these challenges.

CTHNet: a network for wheat ear counting with local-global features fusion based on hybrid architecture. →

Accurate wheat ear counting is one of the key indicators for wheat phenotyping. Convolutional neural network (CNN) algorithms for counting wheat have evolved into sophisticated tools, however because of the limitations of sensory fields, CNN is unable to simulate global context information, which has an impact on counting performance. In this study, we present a hybrid attention network (CTHNet) for wheat ear counting from RGB images that combines local features and global context information. On the one hand, to extract multi-scale local features, a convolutional neural network is built using the Cross Stage Partial framework. On the other hand, to acquire better global context information, tokenized image patches from convolutional neural network feature maps are encoded as input sequences using Pyramid Pooling Transformer. Then, the feature fusion module merges the local features with the global context information to significantly enhance the feature representation. The Global Wheat Head Detection Dataset and Wheat Ear Detection Dataset are used to assess the proposed model. There were 3.40 and 5.21 average absolute errors, respectively. The performance of the proposed model was significantly better than previous studies.

Why it matches plant phenotyping methods小麦穂数という植物形質をRGB画像から推定する深層学習手法を開発し、複数データセットで性能評価しており、表現型取得・抽出法が中心である。

abstractAccurate wheat ear counting is one of the key indicators for wheat phenotyping.

Corn leaf disease: insightful diagnosis using VGG16 empowered by explainable AI. →

The agricultural sector is pivotal to food security and economic stability worldwide. Corn holds particular significance in the global food industry, especially in developing countries where agriculture is a cornerstone of the economy. However, corn crops are vulnerable to various diseases that can significantly reduce yields. Early detection and precise classification of these diseases are crucial to prevent damage and ensure high crop productivity. This study leverages the VGG16 deep learning (DL) model to classify corn leaves into four categories: healthy, blight, gray spot, and common rust. Despite the efficacy of DL models, they often face challenges related to the explainability of their decision-making processes. To address this, Layer-wise Relevance Propagation (LRP) is employed to enhance the model's transparency by generating intuitive and human-readable heat maps of input images. The proposed VGG16 model, augmented with LRP, outperformed previous state-of-the-art models in classifying corn leaf diseases. Simulation results demonstrated that the model not only achieved high accuracy but also provided interpretable results, highlighting critical regions in the images used for classification. By generating human-readable explanations, this approach ensures greater transparency and reliability in model performance, aiding farmers in improving their crop yields.

Why it matches plant phenotyping methodsトウモロコシ葉の画像から病害状態を分類し、VGG16とLRPによる説明可能な解析手法を中心的に開発・評価しているため、植物病害フェノタイピング手法に該当する。

abstractThis study leverages the VGG16 deep learning (DL) model to classify corn leaves into four categories: healthy, blight, gray spot, and common rust.

Semantic segmentation of microbial alterations based on SegFormer. →

Introduction Precise semantic segmentation of microbial alterations is paramount for their evaluation and treatment. This study focuses on harnessing the SegFormer segmentation model for precise semantic segmentation of strawberry diseases, aiming to improve disease detection accuracy under natural acquisition conditions. Methods Three distinct Mix Transformer encoders - MiT-B0, MiT-B3, and MiT-B5 - were thoroughly analyzed to enhance disease detection, targeting diseases such as Angular leaf spot, Anthracnose rot, Blossom blight, Gray mold, Leaf spot, Powdery mildew on fruit, and Powdery mildew on leaves. The dataset consisted of 2,450 raw images, expanded to 4,574 augmented images. The Segment Anything Model integrated into the Roboflow annotation tool facilitated efficient annotation and dataset preparation. Results The results reveal that MiT-B0 demonstrates balanced but slightly overfitting behavior, MiT-B3 adapts rapidly with consistent training and validation performance, and MiT-B5 offers efficient learning with occasional fluctuations, providing robust performance. MiT-B3 and MiT-B5 consistently outperformed MiT-B0 across disease types, with MiT-B5 achieving the most precise segmentation in general. Discussion The findings provide key insights for researchers to select the most suitable encoder for disease detection applications, propelling the field forward for further investigation. The success in strawberry disease analysis suggests potential for extending this approach to other crops and diseases, paving the way for future research and interdisciplinary collaboration.

Why it matches plant phenotyping methodsイチゴ病害の画像から病斑・病害状態をセグメンテーションする手法を開発・比較しており、植物の病害表現型の取得が中心である。

abstractMiT-B0, MiT-B3, and MiT-B5 - were thoroughly analyzed to enhance disease detection

Investigating the water availability hypothesis of pot binding: small pots and infrequent irrigation confound the effects of drought stress in potato ( Solanum tuberosum L.). →

To maximise the throughput of novel, high-throughput phenotyping platforms, many researchers have utilised smaller pot sizes to increase the number of biological replicates that can be grown in spatially limited controlled environments. This may confound plant development through a process known as “pot binding”, particularly in larger species including potato (Solanum tuberosum), and under water-restricted conditions. We aimed to investigate the water availability hypothesis of pot binding, which predicts that small pots have insufficient water holding capacities to prevent drought stress between irrigation periods, in potato. Two cultivars of potato were grown in small (5 L) and large (20 L) pots, were kept under polytunnel conditions, and were subjected to three irrigation frequencies: every other day, daily, and twice daily. Plants were phenotyped with two Phenospex PlantEye F500s and canopy and tuber fresh mass and dry matter were measured. Increasing irrigation frequency from every other day to daily was associated with a significant increase in fresh tuber yield, but only in large pots. This suggests a similar level of drought stress occurred between these treatments in the small pots, supporting the water availability hypothesis of pot binding. Further increasing irrigation frequency to twice daily was still not sufficient to increase yields in small pots but it caused an insignificant increase in yield in the larger pots, suggesting some pot binding may be occurring in large pots under daily irrigation. Canopy temperatures were significantly higher under each irrigation frequency in the small pots compared to large pots, which strongly supports the water availability hypothesis as higher canopy temperatures are a reliable indicator of drought stress in potato. Digital phenotyping was found to be less accurate for larger plants, probably due to a higher degree of self-shading. The research demonstrates the need to define the optimum pot size and irrigation protocols required to completely prevent pot binding and ensure drought treatments are not inadvertently applied to control plants.

Why it matches plant phenotyping methodsPlantEyeを用いたデジタルフェノタイピングの適用と精度評価が研究上の主要要素であり、植物のキャノピー温度や成長状態を測定し、植物サイズによる測定精度低下も検討している。

abstractPlants were phenotyped with two Phenospex PlantEye F500s and canopy and tuber fresh mass and dry matter were measured.

GranoScan: an AI-powered mobile app for in-field identification of biotic threats of wheat. →

Capitalizing on the widespread adoption of smartphones among farmers and the application of artificial intelligence in computer vision, a variety of mobile applications have recently emerged in the agricultural domain. This paper introduces GranoScan, a freely available mobile app accessible on major online platforms, specifically designed for the real-time detection and identification of over 80 threats affecting wheat in the Mediterranean region. Developed through a co-design methodology involving direct collaboration with Italian farmers, this participatory approach resulted in an app featuring: (i) a graphical interface optimized for diverse in-field lighting conditions, (ii) a user-friendly interface allowing swift selection from a predefined menu, (iii) operability even in low or no connectivity, (iv) a straightforward operational guide, and (v) the ability to specify an area of interest in the photo for targeted threat identification. Underpinning GranoScan is a deep learning architecture named efficient minimal adaptive ensembling that was used to obtain accurate and robust artificial intelligence models. The method is based on an ensembling strategy that uses as core models two instances of the EfficientNet-b0 architecture, selected through the weighted F1-score. In this phase a very good precision is reached with peaks of 100% for pests, as well as in leaf damage and root disease tasks, and in some classes of spike and stem disease tasks. For weeds in the post-germination phase, the precision values range between 80% and 100%, while 100% is reached in all the classes for pre-flowering weeds, except one. Regarding recognition accuracy towards end-users in-field photos, GranoScan achieved good performances, with a mean accuracy of 77% and 95% for leaf diseases and for spike, stem and root diseases, respectively. Pests gained an accuracy of up to 94%, while for weeds the app shows a great ability (100% accuracy) in recognizing whether the target weed is a dicot or monocot and 60% accuracy for distinguishing species in both the post-germination and pre-flowering stage. Our precision and accuracy results conform to or outperform those of other studies deploying artificial intelligence models on mobile devices, confirming that GranoScan is a valuable tool also in challenging outdoor conditions.

Why it matches plant phenotyping methods小麦の葉・穂・茎・根の病害や損傷を画像から認識するAIモバイルアプリの開発・性能評価が中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。害虫・雑草識別も含むが、病害認識の技術的評価が明示されている。

abstractThis paper introduces GranoScan, a freely available mobile app accessible on major online platforms, specifically designed for the real-time detection and identification of over 80 threats affecting wheat in the Mediterranean region.

Thermal imaging can reveal variation in stay-green functionality of wheat canopies under temperate conditions. →

Canopy temperature (CT) is often interpreted as representing leaf activity traits such as photosynthetic rates, gas exchange rates, or stomatal conductance. This interpretation is based on the observation that leaf activity traits correlate with transpiration which affects leaf temperature. Accordingly, CT measurements may provide a basis for high throughput assessments of the productivity of wheat canopies during early grain filling, which would allow distinguishing functional from dysfunctional stay-green. However, whereas the usefulness of CT as a fast surrogate measure of sustained vigor under soil drying is well established, its potential to quantify leaf activity traits under high-yielding conditions is less clear. To better understand sensitivity limits of CT measurements under high yielding conditions, we generated within-genotype variability in stay-green functionality by means of differential short-term pre-anthesis canopy shading that modified the sink:source balance. We quantified the effects of these modifications on stay-green properties through a combination of gold standard physiological measurements of leaf activity and newly developed methods for organ-level senescence monitoring based on timeseries of high-resolution imagery and deep-learning-based semantic image segmentation. In parallel, we monitored CT by means of a pole-mounted thermal camera that delivered continuous, ultra-high temporal resolution CT data. Our results show that differences in stay-green functionality translate into measurable differences in CT in the absence of major confounding factors. Differences amounted to approximately 0.8°C and 1.5°C for a very high-yielding source-limited genotype, and a medium-yielding sink-limited genotype, respectively. The gradual nature of the effects of shading on CT during the stay-green phase underscore the importance of a high measurement frequency and a time-integrated analysis of CT, whilst modest effect sizes confirm the importance of restricting screenings to a limited range of morphological and phenological diversity.

Why it matches plant phenotyping methods高解像度画像・深層学習による器官レベル老化モニタリングと熱画像による連続的なキャノピー温度測定を開発・適用し、stay-green機能の表現型評価法として検証しているため、方法が中心的である。

abstractnewly developed methods for organ-level senescence monitoring based on timeseries of high-resolution imagery and deep-learning-based semantic image segmentation

MCCM: multi-scale feature extraction network for disease classification and recognition of chili leaves. →

Currently, foliar diseases of chili have significantly impacted both yield and quality. Despite effective advancements in deep learning techniques for the classification of chili leaf diseases, most existing classification models still face challenges in terms of accuracy and practical application in disease identification. Therefore, in this study, an optimized and enhanced convolutional neural network model named MCCM (MCSAM-ConvNeXt-MSFFM) is proposed by introducing ConvNeXt. The model incorporates a Multi-Scale Feature Fusion Module (MSFFM) aimed at better capturing disease features of various sizes and positions within the images. Moreover, adjustments are made to the positioning, activation functions, and normalization operations of the MSFFM module to further optimize the overall model. Additionally, a proposed Mixed Channel Spatial Attention Mechanism (MCSAM) strengthens the correlation between non-local channels and spatial features, enhancing the model's extraction of fundamental characteristics of chili leaf diseases. During the training process, pre-trained weights are obtained from the Plant Village dataset using transfer learning to accelerate the model's convergence. Regarding model evaluation, the MCCM model is compared with existing CNN models (Vgg16, ResNet34, GoogLeNet, MobileNetV2, ShuffleNet, EfficientNetV2, ConvNeXt), and Swin-Transformer. The results demonstrate that the MCCM model achieves average improvements of 3.38%, 2.62%, 2.48%, and 2.53% in accuracy, precision, recall, and F1 score, respectively. Particularly noteworthy is that compared to the original ConvNeXt model, the MCCM model exhibits significant enhancements across all performance metrics. Furthermore, classification experiments conducted on rice and maize disease datasets showcase the MCCM model's strong generalization performance. Finally, in terms of application, a chili leaf disease classification website is successfully developed using the Flask framework. This website accurately identifies uploaded chili leaf disease images, demonstrating the practical utility of the model.

Why it matches plant phenotyping methodsトウガラシ葉の病徴を画像から分類する深層学習手法を開発・比較評価しており、植物の病害状態を推定するフェノタイピング手法が中心である。

abstractan optimized and enhanced convolutional neural network model named MCCM (MCSAM-ConvNeXt-MSFFM) is proposed

Longitudinal genome-wide association study reveals early QTL that predict biomass accumulation under cold stress in sorghum. →

Introduction: is a promising cellulosic feedstock crop for bioenergy due to its high biomass yields. However, early growth phases of sorghum are sensitive to cold stress, limiting its planting in temperate environments. Cold adaptability is crucial for cultivating bioenergy and grain sorghum at higher latitudes and elevations, or for extending the growing season. Identifying genes and alleles that enhance biomass accumulation under early cold stress can lead to improved sorghum varieties through breeding or genetic engineering. Methods: We conducted image-based phenotyping on 369 accessions from the sorghum Bioenergy Association Panel (BAP) in a controlled environment with early cold treatment. The BAP includes diverse accessions with dense genotyping and varied racial, geographical, and phenotypic backgrounds. Daily, non-destructive imaging allowed temporal analysis of growth-related traits and water use efficiency (WUE). A genome-wide association study (GWAS) was performed to identify genomic intervals and genes associated with cold stress response. Results: The GWAS identified transient quantitative trait loci (QTL) strongly associated with growth-related traits, enabling an exploration of the genetic basis of cold stress response at different developmental stages. This analysis of daily growth traits, rather than endpoint traits, revealed early transient QTL predictive of final phenotypes. The study identified both known and novel candidate genes associated with growth-related traits and temporal responses to cold stress. Discussion: The identified QTL and candidate genes contribute to understanding the genetic mechanisms underlying sorghum's response to cold stress. These findings can inform breeding and genetic engineering strategies to develop sorghum varieties with improved biomass yields and resilience to cold, facilitating earlier planting, extended growing seasons, and cultivation at higher latitudes and elevations.

Why it matches plant phenotyping methods日次の非破壊画像計測を用いて成長関連形質とWUEを時系列で抽出し、早期表現型を解析しており、画像ベースの植物表現型取得が研究の主要な方法として記述されている。

abstractWe conducted image-based phenotyping on 369 accessions from the sorghum Bioenergy Association Panel (BAP) in a controlled environment with early cold treatment.

Detection of maize stem diameter by using RGB-D cameras’ depth information under selected field condition →

Stem diameter is a critical phenotypic parameter for maize, integral to yield prediction and lodging resistance assessment. Traditionally, the quantification of this parameter through manual measurement has been the norm, notwithstanding its tedious and laborious nature. To address these challenges, this study introduces a non-invasive field-based system utilizing depth information from RGB-D cameras to measure maize stem diameter. This technology offers a practical solution for conducting rapid and non-destructive phenotyping. Firstly, RGB images, depth images, and 3D point clouds of maize stems were captured using an RGB-D camera, and precise alignment between the RGB and depth images was achieved. Subsequently, the contours of maize stems were delineated using 2D image processing techniques, followed by the extraction of the stem's skeletal structure employing a thinning-based skeletonization algorithm. Furthermore, within the areas of interest on the maize stems, horizontal lines were constructed using points on the skeletal structure, resulting in 2D pixel coordinates at the intersections of these horizontal lines with the maize stem contours. Subsequently, a back-projection transformation from 2D pixel coordinates to 3D world coordinates was achieved by combining the depth data with the camera's intrinsic parameters. The 3D world coordinates were then precisely mapped onto the 3D point cloud using rigid transformation techniques. Finally, the maize stem diameter was sensed and determined by calculating the Euclidean distance between pairs of 3D world coordinate points. The method demonstrated a Mean Absolute Percentage Error ( MAPE ) of 3.01%, a Mean Absolute Error ( MAE ) of 0.75 mm, a Root Mean Square Error ( RMSE ) of 1.07 mm, and a coefficient of determination ( R ²) of 0.96, ensuring accurate measurement of maize stem diameter. This research not only provides a new method of precise and efficient crop phenotypic analysis but also offers theoretical knowledge for the advancement of precision agriculture.

Why it matches plant phenotyping methodsRGB-Dカメラと画像・3D処理によりトウモロコシ茎径を非破壊測定する手法を開発し、誤差指標で精度検証しており、フェノタイピング手法が中心である。

abstractthis study introduces a non-invasive field-based system utilizing depth information from RGB-D cameras to measure maize stem diameter

Stress phenotyping analysis leveraging autofluorescence image sequences with machine learning. →

Background Autofluorescence-based imaging has the potential to non-destructively characterize the biochemical and physiological properties of plants regulated by genotypes using optical properties of the tissue. A comparative study of stress tolerant and stress susceptible genotypes of Brassica rapa with respect to newly introduced stress-based phenotypes using machine learning techniques will contribute to the significant advancement of autofluorescence-based plant phenotyping research. Methods Autofluorescence spectral images have been used to design a stress detection classifier with two classes, stressed and non-stressed, using machine learning algorithms. The benchmark dataset consisted of time-series image sequences from three Brassica rapa genotypes (CC, R500, and VT), extreme in their morphological and physiological traits captured at the high-throughput plant phenotyping facility at the University of Nebraska-Lincoln, USA. We developed a set of machine learning-based classification models to detect the percentage of stressed tissue derived from plant images and identified the best classifier. From the analysis of the autofluorescence images, two novel stress-based image phenotypes were computed to determine the temporal variation in stressed tissue under progressive drought across different genotypes, i.e., the average percentage stress and the moving average percentage stress. Results The study demonstrated that both the computed phenotypes consistently discriminated against stressed versus non-stressed tissue, with oilseed type (R500) being less prone to drought stress relative to the other two Brassica rapa genotypes (CC and VT). Conclusion Autofluorescence signals from the 365/400 nm excitation/emission combination were able to segregate genotypic variation during a progressive drought treatment under a controlled greenhouse environment, allowing for the exploration of other meaningful phenotypes using autofluorescence image sequences with significance in the context of plant science.

Why it matches plant phenotyping methods自家蛍光画像と機械学習により植物のストレス組織割合を抽出し、新規な時系列ストレス表現型を算出する方法が研究の中心である。

abstractWe developed a set of machine learning-based classification models to detect the percentage of stressed tissue derived from plant images and identified the best classifier.

Tomato leaf disease detection based on attention mechanism and multi-scale feature fusion. →

When detecting tomato leaf diseases in natural environments, factors such as changes in lighting, occlusion, and the small size of leaf lesions pose challenges to detection accuracy. Therefore, this study proposes a tomato leaf disease detection method based on attention mechanisms and multi-scale feature fusion. Firstly, the Convolutional Block Attention Module (CBAM) is introduced into the backbone feature extraction network to enhance the ability to extract lesion features and suppress the effects of environmental interference. Secondly, shallow feature maps are introduced into the re-parameterized generalized feature pyramid network (RepGFPN), constructing a new multi-scale re-parameterized generalized feature fusion module (BiRepGFPN) to enhance feature fusion expression and improve the localization ability for small lesion features. Finally, the BiRepGFPN replaces the Path Aggregation Feature Pyramid Network (PAFPN) in the YOLOv6 model to achieve effective fusion of deep semantic and shallow spatial information. Experimental results indicate that, when evaluated on the publicly available PlantDoc dataset, the model's mean average precision (mAP) showed improvements of 7.7%, 11.8%, 3.4%, 5.7%, 4.3%, and 2.6% compared to YOLOX, YOLOv5, YOLOv6, YOLOv6-s, YOLOv7, and YOLOv8, respectively. When evaluated on the tomato leaf disease dataset, the model demonstrated a precision of 92.9%, a recall rate of 95.2%, an F1 score of 94.0%, and a mean average precision (mAP) of 93.8%, showing improvements of 2.3%, 4.0%, 3.1%, and 2.7% respectively compared to the baseline model. These results indicate that the proposed detection method possesses significant detection performance and generalization capabilities.

Why it matches plant phenotyping methodsトマト葉の病変を画像から検出・局在化する新規深層学習手法を開発し、複数データセットと既存モデルで性能評価しており、植物病害状態の表現型取得が中心である。

abstractthis study proposes a tomato leaf disease detection method based on attention mechanisms and multi-scale feature fusion.

A few-shot learning method for tobacco abnormality identification. →

Tobacco is a valuable crop, but its disease identification is rarely involved in existing works. In this work, we use few-shot learning (FSL) to identify abnormalities in tobacco. FSL is a solution for the data deficiency that has been an obstacle to using deep learning. However, weak feature representation caused by limited data is still a challenging issue in FSL. The weak feature representation leads to weak generalization and troubles in cross-domain. In this work, we propose a feature representation enhancement network (FREN) that enhances the feature representation through instance embedding and task adaptation. For instance embedding, global max pooling, and global average pooling are used together for adding more features, and Gaussian-like calibration is used for normalizing the feature distribution. For task adaptation, self-attention is adopted for task contextualization. Given the absence of publicly available data on tobacco, we created a tobacco leaf abnormality dataset (TLA), which includes 16 categories, two settings, and 1,430 images in total. In experiments, we use PlantVillage, which is the benchmark dataset for plant disease identification, to validate the superiority of FREN first. Subsequently, we use the proposed method and TLA to analyze and discuss the abnormality identification of tobacco. For the multi-symptom diseases that always have low accuracy, we propose a solution by dividing the samples into subcategories created by symptom. For the 10 categories of tomato in PlantVillage, the accuracy achieves 66.04% in 5-way, 1-shot tasks. For the two settings of the tobacco leaf abnormality dataset, the accuracies were achieved at 45.5% and 56.5%. By using the multisymptom solution, the best accuracy can be lifted to 60.7% in 16-way, 1-shot tasks and achieved at 81.8% in 16-way, 10-shot tasks. The results show that our method improves the performance greatly by enhancing feature representation, especially for tasks that contain categories with high similarity. The desensitization of data when crossing domains also validates that the FREN has a strong generalization ability.

Why it matches plant phenotyping methods植物葉の異常・病徴を画像から識別する few-shot 学習法を開発し、専用データセットを作成・検証しており、植物状態の取得・推定手法が中心である。

abstractIn this work, we use few-shot learning (FSL) to identify abnormalities in tobacco.

Genomic prediction within and across maize landrace derived populations using haplotypes. →

Genomic prediction (GP) using haplotypes is considered advantageous compared to GP solely reliant on single nucleotide polymorphisms (SNPs), owing to haplotypes' enhanced ability to capture ancestral information and their higher linkage disequilibrium with quantitative trait loci (QTL). Many empirical studies supported the advantages of haplotype-based GP over SNP-based approaches. Nevertheless, the performance of haplotype-based GP can vary significantly depending on multiple factors, including the traits being studied, the genetic structure of the population under investigation, and the particular method employed for haplotype construction. In this study, we compared haplotype and SNP based prediction accuracies in four populations derived from European maize landraces. Populations comprised either doubled haploid lines (DH) derived directly from landraces, or gamete capture lines (GC) derived from crosses of the landraces with an inbred line. For two different landraces, both types of populations were generated, genotyped with 600k SNPs and phenotyped as lines per se for five traits. Our study explores three prediction scenarios: (i) within each of the four populations, (ii) across DH and GC populations from the same landrace, and (iii) across landraces using either DH or GC populations. Three haplotype construction methods were evaluated: 1. fixed-window blocks (FixedHB), 2. LD-based blocks (HaploView), and 3. IBD-based blocks (HaploBlocker). In within population predictions, FixedHB and HaploView methods performed as well as or slightly better than SNPs for all traits. HaploBlocker improved accuracy for certain traits but exhibited inferior performance for others. In prediction across populations, the parameter setting from HaploBlocker which controls the construction of shared haplotypes between populations played a crucial role for obtaining optimal results. When predicting across landraces, accuracies were low for both, SNP and haplotype approaches, but for specific traits substantial improvement was observed with HaploBlocker. This study provides recommendations for optimal haplotype construction and identifies relevant parameters for constructing haplotypes in the context of genomic prediction.

Why it matches plant phenotyping methodsハプロタイプ構築法を用いたゲノム予測の精度比較・評価が研究の中心であり、複数のトウモロコシ形質の予測という計算的な形質推定法を検証している。

abstractIn this study, we compared haplotype and SNP based prediction accuracies in four populations derived from European maize landraces.

RAAWC-UNet: an apple leaf and disease segmentation method based on residual attention and atrous spatial pyramid pooling improved UNet with weight compression loss. →

Introduction Early detection of leaf diseases is necessary to control the spread of plant diseases, and one of the important steps is the segmentation of leaf and disease images. The uneven light and leaf overlap in complex situations make segmentation of leaves and diseases quite difficult. Moreover, the significant differences in ratios of leaf and disease pixels results in a challenge in identifying diseases. Methods To solve the above issues, the residual attention mechanism combined with atrous spatial pyramid pooling and weight compression loss of UNet is proposed, which is named RAAWC-UNet. Firstly, weights compression loss is a method that introduces a modulation factor in front of the cross-entropy loss, aiming at solving the problem of the imbalance between foreground and background pixels. Secondly, the residual network and the convolutional block attention module are combined to form Res_CBAM. It can accurately localize pixels at the edge of the disease and alleviate the vanishing of gradient and semantic information from downsampling. Finally, in the last layer of downsampling, the atrous spatial pyramid pooling is used instead of two convolutions to solve the problem of insufficient spatial context information. Results The experimental results show that the proposed RAAWC-UNet increases the intersection over union in leaf and disease segmentation by 1.91% and 5.61%, and the pixel accuracy of disease by 4.65% compared with UNet. Discussion The effectiveness of the proposed method was further verified by the better results in comparison with deep learning methods with similar network architectures.

Why it matches plant phenotyping methodsリンゴ葉および病斑の画像セグメンテーション手法を開発し、既存手法と比較検証しているため、植物病害状態の画像ベース表現型計測が中心である。

titleRAAWC-UNet: an apple leaf and disease segmentation method based on residual attention and atrous spatial pyramid pooling improved UNet with weight compression loss.

Seed shape and size of Silene latifolia , differences between sexes, and influence of the parental genome in hybrids with Silene dioica . →

Introduction Plants undergo various natural changes that dramatically modify their genomes. One is polyploidization and the second is hybridization. Both are regarded as key factors in plant evolution and result in phenotypic differences in different plant organs. In Silene , we can find both examples in nature, and this genus has a seed shape diversity that has long been recognized as a valuable source of information for infrageneric classification. Methods Morphometric analysis is a statistical study of shape and size and their covariations with other variables. Traditionally, seed shape description was limited to an approximate comparison with geometric figures (rounded, globular, reniform, or heart-shaped). Seed shape quantification has been based on direct measurements, such as area, perimeter, length, and width, narrowing statistical analysis. We used seed images and processed them to obtain silhouettes. We performed geometric morphometric analyses, such as similarity to geometric models and elliptic Fourier analysis, to study the hybrid offspring of S. latifolia and S. dioica . Results We generated synthetic tetraploids of Silene latifolia and performed controlled crosses between diploid S. latifolia and Silene dioica to analyze seed morphology. After imaging capture and post-processing, statistical analysis revealed differences in seed size, but not in shape, between S. latifolia diploids and tetraploids, as well as some differences in shape among the parentals and hybrids. A detailed inspection using fluorescence microscopy allowed for the identification of shape differences in the cells of the seed coat. In the case of hybrids, differences were found in circularity and solidity. Overal seed shape is maternally regulated for both species, whereas cell shape cannot be associated with any of the sexes. Discussion Our results provide additional tools useful for the combination of morphology with genetics, ecology or taxonomy. Seed shape is a robust indicator that can be used as a complementary tool for the genetic and phylogenetic analyses of Silene hybrid populations.

Why it matches plant phenotyping methods種子画像を処理し、幾何学的形態計測と楕円フーリエ解析で種子形状・サイズを定量化する手法が研究の中心であり、植物器官の形態表現型を抽出している。

abstractWe used seed images and processed them to obtain silhouettes. We performed geometric morphometric analyses, such as similarity to geometric models and elliptic Fourier analysis, to study the hybrid offspring of S. latifolia and S. dioica .

High-throughput UAV-based rice panicle detection and genetic mapping of heading-date-related traits. →

Introduction: ) serves as a vital staple crop that feeds over half the world's population. Optimizing rice breeding for increasing grain yield is critical for global food security. Heading-date-related or Flowering-time-related traits, is a key factor determining yield potential. However, traditional manual phenotyping methods for these traits are time-consuming and labor-intensive. Method: Here we show that aerial imagery from unmanned aerial vehicles (UAVs), when combined with deep learning-based panicle detection, enables high-throughput phenotyping of heading-date-related traits. We systematically evaluated various state-of-the-art object detectors on rice panicle counting and identified YOLOv8-X as the optimal detector. Results: Applying YOLOv8-X to UAV time-series images of 294 rice recombinant inbred lines (RILs) allowed accurate quantification of six heading-date-related traits. Utilizing these phenotypes, we identified quantitative trait loci (QTL), including verified loci and novel loci, associated with heading date. Discussion: Our optimized UAV phenotyping and computer vision pipeline may facilitate scalable molecular identification of heading-date-related genes and guide enhancements in rice yield and adaptation.

Why it matches plant phenotyping methodsUAV画像と深層学習によるイネ穂検出・計数を中核とし、出穂関連形質を高スループットに定量するフェノタイピング手法およびワークフローを評価・適用している。

abstractaerial imagery from unmanned aerial vehicles (UAVs), when combined with deep learning-based panicle detection, enables high-throughput phenotyping of heading-date-related traits

AI-assisted image analysis and physiological validation for progressive drought detection in a diverse panel of Gossypium hirsutum L. →

Introduction Drought detection, spanning from early stress to severe conditions, plays a crucial role in maintaining productivity, facilitating recovery, and preventing plant mortality. While handheld thermal cameras have been widely employed to track changes in leaf water content and stomatal conductance, research on thermal image classification remains limited due mainly to low resolution and blurry images produced by handheld cameras. Methods In this study, we introduce a computer vision pipeline to enhance the significance of leaf-level thermal images across 27 distinct cotton genotypes cultivated in a greenhouse under progressive drought conditions. Our approach involved employing a customized software pipeline to process raw thermal images, generating leaf masks, and extracting a range of statistically relevant thermal features (e.g., min and max temperature, median value, quartiles, etc.). These features were then utilized to develop machine learning algorithms capable of assessing leaf hydration status and distinguishing between well-watered (WW) and dry-down (DD) conditions. Results Two different classifiers were trained to predict the plant treatment-random forest and multilayer perceptron neural networks-finding 75% and 78% accuracy in the treatment prediction, respectively. Furthermore, we evaluated the predicted versus true labels based on classic physiological indicators of drought in plants, including volumetric soil water content, leaf water potential, and chlorophyll a fluorescence, to provide more insights and possible explanations about the classification outputs. Discussion Interestingly, mislabeled leaves mostly exhibited notable responses in fluorescence, water uptake from the soil, and/or leaf hydration status. Our findings emphasize the potential of AI-assisted thermal image analysis in enhancing the informative value of common heterogeneous datasets for drought detection. This application suggests widening the experimental settings to be used with deep learning models, designing future investigations into the genotypic variation in plant drought response and potential optimization of water management in agricultural settings.

Why it matches plant phenotyping methods葉の熱画像からマスクと熱特徴量を抽出し、機械学習で水分状態・乾燥処理を判定する画像解析パイプラインが中心であり、植物表現型の取得・推定手法に該当する。

abstractOur approach involved employing a customized software pipeline to process raw thermal images, generating leaf masks, and extracting a range of statistically relevant thermal features

An RGB image dataset for seed germination prediction and vigor detection - maize. →

Chengcheng Chen1*Muyao Bai1Tairan Wang1Weijia Zhang1Helong Yu2*Tiantian Pang3Jiehong Wu1Zhaokui Li1Xianchang Wang1,3,4

Why it matches plant phenotyping methodsトウモロコシ種子の発芽・活力をRGB画像で推定するデータセットであり、植物表現型の画像取得・解析基盤が中心と判断できる。

titleAn RGB image dataset for seed germination prediction and vigor detection - maize.

Identification of cotton pest and disease based on CFNet- VoV-GCSP -LSKNet-YOLOv8s: a new era of precision agriculture. →

Introduction The study addresses challenges in detecting cotton leaf pests and diseases under natural conditions. Traditional methods face difficulties in this context, highlighting the need for improved identification techniques. Methods The proposed method involves a new model named CFNet-VoV-GCSP-LSKNet-YOLOv8s. This model is an enhancement of YOLOv8s and includes several key modifications: (1) CFNet Module. Replaces all C2F modules in the backbone network to improve multi-scale object feature fusion. (2) VoV-GCSP Module. Replaces C2F modules in the YOLOv8s head, balancing model accuracy with reduced computational load. (3) LSKNet Attention Mechanism. Integrated into the small object layers of both the backbone and head to enhance detection of small objects. (4) XIoU Loss Function. Introduced to improve the model's convergence performance. Results The proposed method achieves high performance metrics: Precision (P), 89.9%. Recall Rate (R), 90.7%. Mean Average Precision (mAP@0.5), 93.7%. The model has a memory footprint of 23.3MB and a detection time of 8.01ms. When compared with other models like YOLO v5s, YOLOX, YOLO v7, Faster R-CNN, YOLOv8n, YOLOv7-tiny, CenterNet, EfficientDet, and YOLOv8s, it shows an average accuracy improvement ranging from 1.2% to 21.8%. Discussion The study demonstrates that the CFNet-VoV-GCSP-LSKNet-YOLOv8s model can effectively identify cotton pests and diseases in complex environments. This method provides a valuable technical resource for the identification and control of cotton pests and diseases, indicating significant improvements over existing methods.

Why it matches plant phenotyping methods綿花葉の病害・害虫を自然条件下の画像から識別する新規YOLOベースモデルを開発し、精度・速度・メモリを比較検証しており、植物の病害状態の取得・推定が中心である。

abstractThe proposed method involves a new model named CFNet-VoV-GCSP-LSKNet-YOLOv8s.

YOLOV5-CBAM-C3TR: an optimized model based on transformer module and attention mechanism for apple leaf disease detection. →

Apple trees face various challenges during cultivation. Apple leaves, as the key part of the apple tree for photosynthesis, occupy most of the area of the tree. Diseases of the leaves can hinder the healthy growth of trees and cause huge economic losses to fruit growers. The prerequisite for precise control of apple leaf diseases is the timely and accurate detection of different diseases on apple leaves. Traditional methods relying on manual detection have problems such as limited accuracy and slow speed. In this study, both the attention mechanism and the module containing the transformer encoder were innovatively introduced into YOLOV5, resulting in YOLOV5-CBAM-C3TR for apple leaf disease detection. The datasets used in this experiment were uniformly RGB images. To better evaluate the effectiveness of YOLOV5-CBAM-C3TR, the model was compared with different target detection models such as SSD, YOLOV3, YOLOV4, and YOLOV5. The results showed that YOLOV5-CBAM-C3TR achieved mAP@0.5, precision, and recall of 73.4%, 70.9%, and 69.5% for three apple leaf diseases including Alternaria blotch, Grey spot, and Rust. Compared with the original model YOLOV5, the mAP 0.5increased by 8.25% with a small change in the number of parameters. In addition, YOLOV5-CBAM-C3TR can achieve an average accuracy of 92.4% in detecting 208 randomly selected apple leaf disease samples. Notably, YOLOV5-CBAM-C3TR achieved 93.1% and 89.6% accuracy in detecting two very similar diseases including Alternaria Blotch and Grey Spot, respectively. The YOLOV5-CBAM-C3TR model proposed in this paper has been applied to the detection of apple leaf diseases for the first time, and also showed strong recognition ability in identifying similar diseases, which is expected to promote the further development of disease detection technology.

Why it matches plant phenotyping methodsリンゴ葉の病徴を画像から検出・分類する新規深層学習モデルを開発し、複数モデルとの比較評価も行っており、植物病害状態の表現型抽出が中心である。

abstractresulting in YOLOV5-CBAM-C3TR for apple leaf disease detection.

Topological data analysis expands the genotype to phenotype map for 3D maize root system architecture →

A central goal of biology is to understand how genetic variation produces phenotypic variation, which has been described as a genotype to phenotype (G to P) map. The plant form is continuously shaped by intrinsic developmental and extrinsic environmental inputs, and therefore plant phenomes are highly multivariate and require comprehensive approaches to fully quantify. Yet a common assumption in plant phenotyping efforts is that a few pre-selected measurements can adequately describe the relevant phenome space. Our poor understanding of the genetic basis of root system architecture is at least partially a result of this incongruence. Root systems are complex 3D structures that are most often studied as 2D representations measured with relatively simple univariate traits. In prior work, we showed that persistent homology, a topological data analysis method that does not pre-suppose the salient features of the data, could expand the phenotypic trait space and identify new G to P relations from a commonly used 2D root phenotyping platform. Here we extend the work to entire 3D root system architectures of maize seedlings from a mapping population that was designed to understand the genetic basis of maize-nitrogen relations. Using a panel of 84 univariate traits, persistent homology methods developed for 3D branching, and multivariate vectors of the collective trait space, we found that each method captures distinct information about root system variation as evidenced by the majority of non-overlapping QTL, and hence that root phenotypic trait space is not easily exhausted. The work offers a data-driven method for assessing 3D root structure and highlights the importance of non-canonical phenotypes for more accurate representations of the G to P map.

Why it matches plant phenotyping methods3Dトウモロコシ根系の構造を、persistent homologyによって従来の形質空間を拡張して解析する計算的フェノタイピング手法が中心であり、根系構造形質の抽出・評価に該当する。

abstractpersistent homology, a topological data analysis method that does not pre-suppose the salient features of the data, could expand the phenotypic trait space

A robust and light-weight transfer learning-based architecture for accurate detection of leaf diseases across multiple plants using less amount of images. →

Leaf diseases are a global threat to crop production and food preservation. Detecting these diseases is crucial for effective management. We introduce LeafDoc-Net, a robust, lightweight transfer-learning architecture for accurately detecting leaf diseases across multiple plant species, even with limited image data. Our approach concatenates two pre-trained image classification deep learning-based models, DenseNet121 and MobileNetV2. We enhance DenseNet121 with an attention-based transition mechanism and global average pooling layers, while MobileNetV2 benefits from adding an attention module and global average pooling layers. We deepen the architecture with extra-dense layers featuring swish activation and batch normalization layers, resulting in a more robust and accurate model for diagnosing leaf-related plant diseases. LeafDoc-Net is evaluated on two distinct datasets, focused on cassava and wheat leaf diseases, demonstrating superior performance compared to existing models in accuracy, precision, recall, and AUC metrics. To gain deeper insights into the model's performance, we utilize Grad-CAM++.

Why it matches plant phenotyping methods植物葉の病徴を画像から分類する深層学習手法を開発し、複数データセットで性能評価しており、植物フェノタイピング手法が研究の中心である。

abstractWe introduce LeafDoc-Net, a robust, lightweight transfer-learning architecture for accurately detecting leaf diseases across multiple plant species, even with limited image data.

Genomic prediction reveals unexplored variation in grain protein and lysine content across a vast winter wheat genebank collection. →

Globally, wheat ( Triticum aestivum L.) is a major source of proteins in human nutrition despite its unbalanced amino acid composition. The low lysine content in the protein fraction of wheat can lead to protein-energy-malnutrition prominently in developing countries. A promising strategy to overcome this problem is to breed varieties which combine high protein content with high lysine content. Nevertheless, this requires the incorporation of yet undefined donor genotypes into pre-breeding programs. Genebank collections are suspected to harbor the needed genetic diversity. In the 1970s, a large-scale screening of protein traits was conducted for the wheat genebank collection in Gatersleben; however, this data has been poorly mined so far. In the present study, a large historical dataset on protein content and lysine content of 4,971 accessions was curated, strictly corrected for outliers as well as for unreplicated data and consolidated as the corresponding adjusted entry means. Four genomic prediction approaches were compared based on the ability to accurately predict the traits of interest. High-quality phenotypic data of 558 accessions was leveraged by engaging the best performing prediction model, namely EG-BLUP. Finally, this publication incorporates predicted phenotypes of 7,651 accessions of the winter wheat collection. Five accessions were proposed as donor genotypes due to the combination of outstanding high protein content as well as lysine content. Further investigation of the passport data suggested an association of the adjusted lysine content with the elevation of the collecting site. This publicly available information can facilitate future pre-breeding activities.

Why it matches plant phenotyping methods小麦のタンパク質・リジン含量という植物形質について、歴史的表現型データを整理し、複数のゲノム予測法を比較して大規模コレクションの予測表現型を生成しており、計算的な形質推定とデータセット活用が研究の中心です。

abstracta large historical dataset on protein content and lysine content of 4,971 accessions was curated, strictly corrected for outliers as well as for unreplicated data and consolidated as the corresponding adjusted entry means.

PDSE-Lite: lightweight framework for plant disease severity estimation based on Convolutional Autoencoder and Few-Shot Learning →

Plant disease diagnosis with estimation of disease severity at early stages still remains a significant research challenge in agriculture. It is helpful in diagnosing plant diseases at the earliest so that timely action can be taken for curing the disease. Existing studies often rely on labor-intensive manually annotated large datasets for disease severity estimation. In order to conquer this problem, a lightweight framework named “PDSE-Lite” based on Convolutional Autoencoder (CAE) and Few-Shot Learning (FSL) is proposed in this manuscript for plant disease severity estimation with few training instances. The PDSE-Lite framework is designed and developed in two stages. In first stage, a lightweight CAE model is built and trained to reconstruct leaf images from original leaf images with minimal reconstruction loss. In subsequent stage, pretrained layers of the CAE model built in the first stage are utilized to develop the image classification and segmentation models, which are then trained using FSL. By leveraging FSL, the proposed framework requires only a few annotated instances for training, which significantly reduces the human efforts required for data annotation. Disease severity is then calculated by determining the percentage of diseased leaf pixels obtained through segmentation out of the total leaf pixels. The PDSE-Lite framework’s performance is evaluated on Apple-Tree-Leaf-Disease-Segmentation (ATLDS) dataset. However, the proposed framework can identify any plant disease and quantify the severity of identified diseases. Experimental results reveal that the PDSE-Lite framework can accurately detect healthy and four types of apple tree diseases as well as precisely segment the diseased area from leaf images by using only two training samples from each class of the ATLDS dataset. Furthermore, the PDSE-Lite framework’s performance is compared with existing state-of-the-art techniques, and it is found that this framework outperformed these approaches. The proposed framework’s applicability is further verified by statistical hypothesis testing using Student t-test. The results obtained from this test confirm that the proposed framework can precisely estimate the plant disease severity with a confidence interval of 99%. Hence, by reducing the reliance on large-scale manual data annotation, the proposed framework offers a promising solution for early-stage plant disease diagnosis and severity estimation.

Why it matches plant phenotyping methods植物葉画像から病変画素率を算出して病害重症度を推定する画像解析手法を開発し、データセット上で比較・統計検証しており、植物表現型取得が中心である。

abstracta lightweight framework named “PDSE-Lite” based on Convolutional Autoencoder (CAE) and Few-Shot Learning (FSL) is proposed in this manuscript for plant disease severity estimation with few training instances.

Deep learning architectures for diagnosing the severity of apple frog-eye leaf spot disease in complex backgrounds. →

Introduction In precision agriculture, accurately diagnosing apple frog-eye leaf spot disease is critical for effective disease management. Traditional methods, predominantly relying on labor-intensive and subjective visual evaluations, are often inefficient and unreliable. Methods To tackle these challenges in complex orchard environments, we develop a specialized deep learning architecture. This architecture consists of a two-stage multi-network model. The first stage features an enhanced Pyramid Scene Parsing Network (L-DPNet) with deformable convolutions for improved apple leaf segmentation. The second stage utilizes an improved U-Net (D-UNet), optimized with bilinear upsampling and batch normalization, for precise disease spot segmentation. Results Our model sets new benchmarks in performance, achieving a mean Intersection over Union (mIoU) of 91.27% for segmentation of both apple leaves and disease spots, and a mean Pixel Accuracy (mPA) of 94.32%. It also excels in classifying disease severity across five levels, achieving an overall precision of 94.81%. Discussion This approach represents a significant advancement in automated disease quantification, enhancing disease management in precision agriculture through data-driven decision-making.

Why it matches plant phenotyping methodsリンゴ葉の病斑を画像から分割し、病害重症度を定量化・分類する深層学習手法が研究の中心であり、植物病態の表現型推定に該当する。

abstractwe develop a specialized deep learning architecture

Zero-exemplar deep continual learning for crop disease recognition: a study of total variation attention regularization in vision transformers →

With the increasing integration of AI technology in the food industry, deep learning has demonstrated its immense potential in the domain of plant disease image recognition. However, there remains a gap in research between models capable of continual learning of new diseases and addressing the inherent catastrophic forgetting issue in neural networks. This study aims to comprehensively evaluate various learning strategies based on advanced computer vision models for multi-disease continual learning tasks in food disease recognition. To cater to the benchmark dataset requirements, we collected the PlantDiseaseCL dataset, sourced from the internet, encompassing diverse crop diseases from apples, corn, and more. Utilizing the Vision Transformer (ViT) model, we established a plant disease image recognition classifier, which, in joint learning, outperformed several comparative CNN architectures in accuracy (0.9538), precision (0.9532), recall (0.9528), and F1 score (0.9560). To further harness the potential of ViT in food disease defect recognition, we introduced a mathematical paradigm for crop disease recognition continual learning. For the first time, we proposed a novel ViT-TV architecture in the multi-disease image recognition scenario, incorporating a Total Variation (TV) distance-based loss (TV-Loss) to quantify the disparity between current and previous attention distributions, fostering attention consistency and mitigating the catastrophic forgetting inherent in ViT without prior task samples. In the incremental learning of the PlantDiseaseCL dataset across 3-Steps and 5-Steps, our strategy achieved average accuracies of 0.7077 and 0.5661, respectively, surpassing all compared Zero-Exemplar Approaches like LUCIR, SI, MAS, and even outperforming exemplar-based strategies like EEIL and ICaRL. In conclusion, the ViT-TV approach offers robust support for the long-term intelligent development of the agricultural and food industry, especially showcasing significant applicability in continual learning for crop disease image recognition.

Why it matches plant phenotyping methods植物病害画像から病害状態を推定する認識手法を開発・評価し、継続学習用データセットも構築しているため、植物フェノタイピング手法が中心である。

abstractwe collected the PlantDiseaseCL dataset, sourced from the internet, encompassing diverse crop diseases from apples, corn, and more.

Estimating the frost damage index in lettuce using UAV-based RGB and multispectral images. →

Introduction The cold stress is one of the most important factors for affecting production throughout year, so effectively evaluating frost damage is great significant to the determination of the frost tolerance in lettuce. Methods We proposed a high-throughput method to estimate lettuce FDI based on remote sensing. Red-Green-Blue (RGB) and multispectral images of open-field lettuce suffered from frost damage were captured by Unmanned Aerial Vehicle platform. Pearson correlation analysis was employed to select FDI-sensitive features from RGB and multispectral images. Then the models were established for different FDI-sensitive features based on sensor types and different groups according to lettuce colors using multiple linear regression, support vector machine and neural network algorithms, respectively. Results and discussion Digital number of blue and red channels, spectral reflectance at blue, red and near-infrared bands as well as six vegetation indexes (VIs) were found to be significantly related to the FDI of all lettuce groups. The high sensitivity of four modified VIs to frost damage of all lettuce groups was confirmed. The average accuracy of models were improved by 3% to 14% through a combination of multisource features. Color of lettuce had a certain impact on the monitoring of frost damage by FDI prediction models, because the accuracy of models based on green lettuce group were generally higher. The MULTISURCE-GREEN-NN model with R 2 of 0.715 and RMSE of 0.014 had the best performance, providing a high-throughput and efficient technical tool for frost damage investigation which will assist the identification of cold-resistant green lettuce germplasm and related breeding.

Why it matches plant phenotyping methodsUAV画像と機械学習を用いてレタスの霜害指数という植物状態を推定する高スループット手法を開発・評価しており、フェノタイピング手法が中心です。

abstractWe proposed a high-throughput method to estimate lettuce FDI based on remote sensing.

Plant disease detection model for edge computing devices. →

In this paper, we address the question of achieving high accuracy in deep learning models for agricultural applications through edge computing devices while considering the associated resource constraints. Traditional and state-of-the-art models have demonstrated good accuracy, but their practicality as end-user available solutions remains uncertain due to current resource limitations. One agricultural application for deep learning models is the detection and classification of plant diseases through image-based crop monitoring. We used the publicly available PlantVillage dataset containing images of healthy and diseased leaves for 14 crop species and 6 groups of diseases as example data. The MobileNetV3-small model succeeds in classifying the leaves with a test accuracy of around 99.50%. Post-training optimization using quantization reduced the number of model parameters from approximately 1.5 million to 0.93 million while maintaining the accuracy of 99.50%. The final model is in ONNX format, enabling deployment across various platforms, including mobile devices. These findings offer a cost-effective solution for deploying accurate deep-learning models in agricultural applications.

Why it matches plant phenotyping methods植物葉画像から病害を分類する深層学習モデルの開発とエッジ展開が中心であり、植物の病害状態を直接推定するフェノタイピング手法に該当する。

abstractthe detection and classification of plant diseases through image-based crop monitoring

Harnessing the power of diffusion models for plant disease image augmentation →

Introduction The challenges associated with data availability, class imbalance, and the need for data augmentation are well-recognized in the field of plant disease detection. The collection of large-scale datasets for plant diseases is particularly demanding due to seasonal and geographical constraints, leading to significant cost and time investments. Traditional data augmentation techniques, such as cropping, resizing, and rotation, have been largely supplanted by more advanced methods. In particular, the utilization of Generative Adversarial Networks (GANs) for the creation of realistic synthetic images has become a focal point of contemporary research, addressing issues related to data scarcity and class imbalance in the training of deep learning models. Recently, the emergence of diffusion models has captivated the scientific community, offering superior and realistic output compared to GANs. Despite these advancements, the application of diffusion models in the domain of plant science remains an unexplored frontier, presenting an opportunity for groundbreaking contributions. Methods In this study, we delve into the principles of diffusion technology, contrasting its methodology and performance with state-of-the-art GAN solutions, specifically examining the guided inference model of GANs, named InstaGAN, and a diffusion-based model, RePaint. Both models utilize segmentation masks to guide the generation process, albeit with distinct principles. For a fair comparison, a subset of the PlantVillage dataset is used, containing two disease classes of tomato leaves and three disease classes of grape leaf diseases, as results on these classes have been published in other publications. Results Quantitatively, RePaint demonstrated superior performance over InstaGAN, with average Fréchet Inception Distance (FID) score of 138.28 and Kernel Inception Distance (KID) score of 0.089 ± (0.002), compared to InstaGAN’s average FID and KID scores of 206.02 and 0.159 ± (0.004) respectively. Additionally, RePaint’s FID scores for grape leaf diseases were 69.05, outperforming other published methods such as DCGAN (309.376), LeafGAN (178.256), and InstaGAN (114.28). For tomato leaf diseases, RePaint achieved an FID score of 161.35, surpassing other methods like WGAN (226.08), SAGAN (229.7233), and InstaGAN (236.61). Discussion This study offers valuable insights into the potential of diffusion models for data augmentation in plant disease detection, paving the way for future research in this promising field.

Why it matches plant phenotyping methods植物病害画像を対象に拡散モデルとGANを比較し、病害画像データ拡張の性能をFID・KIDで検証する研究であり、植物病害状態の画像ベース評価を支える方法が中心である。

titleHarnessing the power of diffusion models for plant disease image augmentation

OSC-CO2: coattention and cosegmentation framework for plant state change with multiple features →

Cosegmentation and coattention are extensions of traditional segmentation methods aimed at detecting a common object (or objects) in a group of images. Current cosegmentation and coattention methods are ineffective for objects, such as plants, that change their morphological state while being captured in different modalities and views. The Object State Change using Coattention-Cosegmentation (OSC-CO2) is an end-to-end unsupervised deep-learning framework that enhances traditional segmentation techniques, processing, analyzing, selecting, and combining suitable segmentation results that may contain most of our target object’s pixels, and then displaying a final segmented image. The framework leverages coattention-based convolutional neural networks (CNNs) and cosegmentation-based dense Conditional Random Fields (CRFs) to address segmentation accuracy in high-dimensional plant imagery with evolving plant objects. The efficacy of OSC-CO2 is demonstrated using plant growth sequences imaged with infrared, visible, and fluorescence cameras in multiple views using a remote sensing, high-throughput phenotyping platform, and is evaluated using Jaccard index and precision measures. We also introduce CosegPP+, a dataset that is structured and can provide quantitative information on the efficacy of our framework. Results show that OSC-CO2 out performed state-of-the art segmentation and cosegmentation methods by improving segementation accuracy by 3% to 45%.

Why it matches plant phenotyping methods植物画像から成長状態を抽出する画像セグメンテーション手法を開発し、ハイスループット表現型解析プラットフォームで評価しているため、方法が研究の中心である。

abstractThe Object State Change using Coattention-Cosegmentation (OSC-CO2) is an end-to-end unsupervised deep-learning framework

MS-Net: a novel lightweight and precise model for plant disease identification →

The rapid development of image processing technology and the improvement of computing power in recent years have made deep learning one of the main methods for plant disease identification. Currently, many neural network models have shown better performance in plant disease identification. Typically, the performance improvement of the model needs to be achieved by increasing the depth of the network. However, this also increases the computational complexity, memory requirements, and training time, which will be detrimental to the deployment of the model on mobile devices. To address this problem, a novel lightweight convolutional neural network has been proposed for plant disease detection. Skip connections are introduced into the conventional MobileNetV3 network to enrich the input features of the deep network, and the feature fusion weight parameters in the skip connections are optimized using an improved whale optimization algorithm to achieve higher classification accuracy. In addition, the bias loss substitutes the conventional cross-entropy loss to reduce the interference caused by redundant data during the learning process. The proposed model is pre-trained on the plant classification task dataset instead of using the classical ImageNet for pre-training, which further enhances the performance and robustness of the model. The constructed network achieved high performance with fewer parameters, reaching an accuracy of 99.8% on the PlantVillage dataset. Encouragingly, it also achieved a prediction accuracy of 97.8% on an apple leaf disease dataset with a complex outdoor background. The experimental results show that compared with existing advanced plant disease diagnosis models, the proposed model has fewer parameters, higher recognition accuracy, and lower complexity.

Why it matches plant phenotyping methods植物病害画像から病害状態を推定する軽量CNNモデルを開発し、複数データセットで精度・複雑度を評価しており、病害表現型の抽出手法が中心である。

abstracta novel lightweight convolutional neural network has been proposed for plant disease detection

Modeling the spatial-spectral characteristics of plants for nutrient status identification using hyperspectral data and deep learning methods. →

Sustainable fertilizer management in precision agriculture is essential for both economic and environmental reasons. To effectively manage fertilizer input, various methods are employed to monitor and track plant nutrient status. One such method is hyperspectral imaging, which has been on the rise in recent times. It is a remote sensing tool used to monitor plant physiological changes in response to environmental conditions and nutrient availability. However, conventional hyperspectral processing mainly focuses on either the spectral or spatial information of plants. This study aims to develop a hybrid convolution neural network (CNN) capable of simultaneously extracting spatial and spectral information from quinoa and cowpea plants to identify their nutrient status at different growth stages. To achieve this, a nutrient experiment with four treatments (high and low levels of nitrogen and phosphorus) was conducted in a glasshouse. A hybrid CNN model comprising a 3D CNN (extracts joint spectral-spatial information) and a 2D CNN (for abstract spatial information extraction) was proposed. Three pre-processing techniques, including second-order derivative, standard normal variate, and linear discriminant analysis, were applied to selected regions of interest within the plant spectral hypercube. Together with the raw data, these datasets were used as inputs to train the proposed model. This was done to assess the impact of different pre-processing techniques on hyperspectral-based nutrient phenotyping. The performance of the proposed model was compared with a 3D CNN, a 2D CNN, and a Hybrid Spectral Network (HybridSN) model. Effective wavebands were selected from the best-performing dataset using a greedy stepwise-based correlation feature selection (CFS) technique. The selected wavebands were then used to retrain the models to identify the nutrient status at five selected plant growth stages. From the results, the proposed hybrid model achieved a classification accuracy of over 94% on the test dataset, demonstrating its potential for identifying nitrogen and phosphorus status in cowpea and quinoa at different growth stages.

Why it matches plant phenotyping methods植物の栄養状態をハイパースペクトル画像から抽出するCNN手法を開発し、前処理・複数モデルとの比較・異なる生育段階での性能評価を行っており、表現型取得が中心である。

abstractThis study aims to develop a hybrid convolution neural network (CNN) capable of simultaneously extracting spatial and spectral information from quinoa and cowpea plants to identify their nutrient status at different growth stages.

Size measurement and filled/unfilled detection of rice grains using backlight image processing. →

Measurements of rice physical traits, such as length, width, and percentage of filled/unfilled grains, are essential steps of rice breeding. A new approach for measuring the physical traits of rice grains for breeding purposes was presented in this study, utilizing image processing techniques. Backlight photography was used to capture a grayscale image of a group of rice grains, which was then analyzed using a clustering algorithm to differentiate between filled and unfilled grains based on their grayscale values. The impact of backlight intensity on the accuracy of the method was also investigated. The results show that the proposed method has excellent accuracy and high efficiency. The mean absolute percentage error of the method was 0.24% and 1.36% in calculating the total number of grain particles and distinguishing the number of filled grains, respectively. The grain size was also measured with a little margin of error. The mean absolute percentage error of grain length measurement was 1.11%, while the measurement error of grain width was 4.03%. The method was found to be highly accurate, non-destructive, and cost-effective when compared to conventional methods, making it a promising approach for characterizing physical traits for crop breeding.

Why it matches plant phenotyping methodsイネ籾の長さ・幅・充実度を画像処理で測定する方法を開発し、精度とバックライト条件の影響を検証しており、植物表現型取得が中心です。

abstractA new approach for measuring the physical traits of rice grains for breeding purposes was presented in this study, utilizing image processing techniques.

TeaDiseaseNet: multi-scale self-attentive tea disease detection. →

Accurate detection of tea diseases is essential for optimizing tea yield and quality, improving production, and minimizing economic losses. In this paper, we introduce TeaDiseaseNet, a novel disease detection method designed to address the challenges in tea disease detection, such as variability in disease scales and dense, obscuring disease patterns. TeaDiseaseNet utilizes a multi-scale self-attention mechanism to enhance disease detection performance. Specifically, it incorporates a CNN-based module for extracting features at multiple scales, effectively capturing localized information such as texture and edges. This approach enables a comprehensive representation of tea images. Additionally, a self-attention module captures global dependencies among pixels, facilitating effective interaction between global information and local features. Furthermore, we integrate a channel attention mechanism, which selectively weighs and combines the multi-scale features, eliminating redundant information and enabling precise localization and recognition of tea disease information across diverse scales and complex backgrounds. Extensive comparative experiments and ablation studies validate the effectiveness of the proposed method, demonstrating superior detection results in scenarios characterized by complex backgrounds and varying disease scales. The presented method provides valuable insights for intelligent tea disease diagnosis, with significant potential for improving tea disease management and production.

Why it matches plant phenotyping methods茶葉画像から病害状態を推定する新規画像解析手法を開発し、比較実験とアブレーションで検証しており、植物病害フェノタイピングが中心です。

abstractwe introduce TeaDiseaseNet, a novel disease detection method designed to address the challenges in tea disease detection

An effective approach for plant leaf diseases classification based on a novel DeepPlantNet deep learning model →

Introduction Recently, plant disease detection and diagnosis procedures have become a primary agricultural concern. Early detection of plant diseases enables farmers to take preventative action, stopping the disease's transmission to other plant sections. Plant diseases are a severe hazard to food safety, but because the essential infrastructure is missing in various places around the globe, quick disease diagnosis is still difficult. The plant may experience a variety of attacks, from minor damage to total devastation, depending on how severe the infections are. Thus, early detection of plant diseases is necessary to optimize output to prevent such destruction. The physical examination of plant diseases produced low accuracy, required a lot of time, and could not accurately anticipate the plant disease. Creating an automated method capable of accurately classifying to deal with these issues is vital. Method This research proposes an efficient, novel, and lightweight DeepPlantNet deep learning (DL)-based architecture for predicting and categorizing plant leaf diseases. The proposed DeepPlantNet model comprises 28 learned layers, i.e., 25 convolutional layers (ConV) and three fully connected (FC) layers. The framework employed Leaky RelU (LReLU), batch normalization (BN), fire modules, and a mix of 3×3 and 1×1 filters, making it a novel plant disease classification framework. The Proposed DeepPlantNet model can categorize plant disease images into many classifications. Results The proposed approach categorizes the plant diseases into the following ten groups: Apple_Black_rot (ABR), Cherry_(including_sour)_Powdery_mildew (CPM), Grape_Leaf_blight_(Isariopsis_Leaf_Spot) (GLB), Peach_Bacterial_spot (PBS), Pepper_bell_Bacterial_spot (PBBS), Potato_Early_blight (PEB), Squash_Powdery_mildew (SPM), Strawberry_Leaf_scorch (SLS), bacterial tomato spot (TBS), and maize common rust (MCR). The proposed framework achieved an average accuracy of 98.49 and 99.85in the case of eight-class and three-class classification schemes, respectively. Discussion The experimental findings demonstrated the DeepPlantNet model's superiority to the alternatives. The proposed technique can reduce financial and agricultural output losses by quickly and effectively assisting professionals and farmers in identifying plant leaf diseases.

Why it matches plant phenotyping methods葉画像から植物病害状態を分類する深層学習手法を開発しており、植物の病徴・病害状態の取得と推定が研究の中心であるため。

abstractThis research proposes an efficient, novel, and lightweight DeepPlantNet deep learning (DL)-based architecture for predicting and categorizing plant leaf diseases.

ShinyFruit: interactive fruit phenotyping software and its application in blackberry →

Introduction: Horticultural plant breeding programs often demand large volumes of phenotypic data to capture visual variation in quality of harvested products. Increasing the throughput potential of phenomic pipelines enables breeders to consider data-hungry molecular breeding strategies such as genome-wide association studies and genomic selection. Methods: We present an R-based web application called ShinyFruit for image-based phenotyping of size, shape, and color-related qualities in fruits and vegetables. Here, we have demonstrated one potential application for ShinyFruit by comparing its estimates of fruit length, width, and red drupelet reversion (RDR) with ImageJ and analogous manual phenotyping techniques in a population of blackberry cultivars and breeding selections from the University of Arkansas System Division of Agriculture Fruit Breeding Program. Results: = 0.62 - 0.70). Neither phenotyping method detected genotypic differences in blackberry fruit width, suggesting that this trait is unlikely to be heritable in the population observed. Discussion: It is likely that implementing a treatment to promote RDR expression in future studies might strengthen the documented correlation between phenotyping methods by maximizing genotypic variance. Even so, our analysis has suggested that ShinyFruit provides a viable, open-source solution to efficient phenotyping of size and color in blackberry fruit. The ability for users to adjust analysis settings should also extend its utility to a wide range of fruits and vegetables.

Why it matches plant phenotyping methodsShinyFruitは果実のサイズ・形状・色を画像から推定するソフトウェアであり、ImageJおよび手動測定との比較検証も行っているため、植物表現型取得法が中心である。

abstractWe present an R-based web application called ShinyFruit for image-based phenotyping of size, shape, and color-related qualities in fruits and vegetables.

A New Deep Learning-based Dynamic Paradigm Towards Open-World Plant Disease Detection. →

Plant disease detection has made significant strides thanks to the emergence of deep learning. However, existing methods have been limited to closed-set and static learning settings, where models are trained using a specific dataset. This confinement restricts the model's adaptability when encountering samples from unseen disease categories. Additionally, there is a challenge of knowledge degradation for these static learning settings, as the acquisition of new knowledge tends to overwrite the old when learning new categories. To overcome these limitations, this study introduces a novel paradigm for plant disease detection called open-world setting. Our approach can infer disease categories that have never been seen during the model training phase and gradually learn these unseen diseases through dynamic knowledge updates in the next training phase. Specifically, we utilize a well-trained unknown-aware region proposal network to generate pseudo-labels for unknown diseases during training and employ a class-agnostic classifier to enhance the recall rate for unknown diseases. Besides, we employ a sample replay strategy to maintain recognition ability for previously learned classes. Extensive experimental evaluation and ablation studies investigate the efficacy of our method in detecting old and unknown classes. Remarkably, our method demonstrates robust generalization ability even in cross-species disease detection experiments. Overall, this open-world and dynamically updated detection method shows promising potential to become the future paradigm for plant disease detection. We discuss open issues including classification and localization, and propose promising approaches to address them. We encourage further research in the community to tackle the crucial challenges in open-world plant disease detection. The code will be released at https://github.com/JiuqingDong/OWPDD.

Why it matches plant phenotyping methods植物の病徴・病害状態を画像から検出する手法の開発と評価が中心であり、未知病害への検出・学習、アブレーションおよび交差種評価を行っているため、植物フェノタイピング手法として採用する。

abstractSpecifically, we utilize a well-trained unknown-aware region proposal network to generate pseudo-labels for unknown diseases during training and employ a class-agnostic classifier to enhance the recall rate for unknown diseases.

Quantifying physiological trait variation with automated hyperspectral imaging in rice. →

Advancements in hyperspectral imaging (HSI) together with the establishment of dedicated plant phenotyping facilities worldwide have enabled high-throughput collection of plant spectral images with the aim of inferring target phenotypes. Here, we test the utility of HSI-derived canopy data, which were collected as part of an automated plant phenotyping system, to predict physiological traits in cultivated Asian rice ( Oryza sativa ). We evaluated 23 genetically diverse rice accessions from two subpopulations under two contrasting nitrogen conditions and measured 14 leaf- and canopy-level parameters to serve as ground-reference observations. HSI-derived data were used to (1) classify treatment groups across multiple vegetative stages using support vector machines (≥ 83% accuracy) and (2) predict leaf-level nitrogen content (N, %, n=88 ) and carbon to nitrogen ratio (C:N, n=88 ) with Partial Least Squares Regression (PLSR) following RReliefF wavelength selection (validation: R 2 = 0.797 and RMSEP = 0.264 for N; R 2 = 0.592 and RMSEP = 1.688 for C:N). Results demonstrated that models developed using training data from one rice subpopulation were able to predict N and C:N in the other subpopulation, while models trained on a single treatment group were not able to predict samples from the other treatment. Finally, optimization of PLSR-RReliefF hyperparameters showed that 300-400 wavelengths generally yielded the best model performance with a minimum calibration sample size of 62. Results support the use of canopy-level hyperspectral imaging data to estimate leaf-level N and C:N across diverse rice, and this work highlights the importance of considering calibration set design prior to data collection as well as hyperparameter optimization for model development in future studies.

Why it matches plant phenotyping methods自動ハイパースペクトル画像からイネの生理形質を推定するモデルを開発・検証しており、表現型取得・抽出手法が研究の中心である。

abstractHSI-derived data were used to (1) classify treatment groups across multiple vegetative stages using support vector machines (≥ 83% accuracy) and (2) predict leaf-level nitrogen content (N, %, n=88 ) and carbon to nitrogen ratio (C:N, n=88 ) with Partial Least Squares Regression (PLSR) following RReliefF wavelength selection

High-throughput and separating-free phenotyping method for on-panicle rice grains based on deep learning →

Rice is a vital food crop that feeds most of the global population. Cultivating high-yielding and superior-quality rice varieties has always been a critical research direction. Rice grain-related traits can be used as crucial phenotypic evidence to assess yield potential and quality. However, the analysis of rice grain traits is still mainly based on manual counting or various seed evaluation devices, which incur high costs in time and money. This study proposed a high-precision phenotyping method for rice panicles based on visible light scanning imaging and deep learning technology, which can achieve high-throughput extraction of critical traits of rice panicles without separating and threshing rice panicles. The imaging of rice panicles was realized through visible light scanning. The grains were detected and segmented using the Faster R-CNN-based model, and an improved Pix2Pix model cascaded with it was used to compensate for the information loss caused by the natural occlusion between the rice grains. An image processing pipeline was designed to calculate fifteen phenotypic traits of the on-panicle rice grains. Eight varieties of rice were used to verify the reliability of this method. The R 2 values between the extraction by the method and manual measurements of the grain number, grain length, grain width, grain length/width ratio and grain perimeter were 0.99, 0.96, 0.83, 0.90 and 0.84, respectively. Their mean absolute percentage error (MAPE) values were 1.65%, 7.15%, 5.76%, 9.13% and 6.51%. The average imaging time of each rice panicle was about 60 seconds, and the total time of data processing and phenotyping traits extraction was less than 10 seconds. By randomly selecting one thousand grains from each of the eight varieties and analyzing traits, it was found that there were certain differences between varieties in the number distribution of thousand-grain length, thousand-grain width, and thousand-grain length/width ratio. The results show that this method is suitable for high-throughput, non-destructive, and high-precision extraction of on-panicle grains traits without separating. Low cost and robust performance make it easy to popularize. The research results will provide new ideas and methods for extracting panicle traits of rice and other crops.

Why it matches plant phenotyping methodsイネ穂上粒の形態形質を画像・深層学習で抽出する手法を開発し、手動測定との比較で検証しており、表現型取得が研究の中心である。

abstractThis study proposed a high-precision phenotyping method for rice panicles based on visible light scanning imaging and deep learning technology

Remote sensing continuity: a comparison of HTP platforms and potential challenges with field applications. →

In an era of climate change and increased environmental variability, breeders are looking for tools to maintain and increase genetic gain and overall efficiency. In recent years the field of high throughput phenotyping (HTP) has received increased attention as an option to meet this need. There are many platform options in HTP, but ground-based handheld and remote aerial systems are two popular options. While many HTP setups have similar specifications, it is not always clear if data from different systems can be treated interchangeably. In this research, we evaluated two handheld radiometer platforms, Cropscan MSR16R and Spectra Vista Corp (SVC) HR-1024i, as well as a UAS-based system with a Sentera Quad Multispectral Sensor. Each handheld radiometer was used for two years simultaneously with the unoccupied aircraft systems (UAS) in collecting winter wheat breeding trials between 2018-2021. Spectral reflectance indices (SRI) were calculated for each system. SRI heritability and correlation were analyzed in evaluating the platform and SRI usability for breeding applications. Correlations of SRIs were low against UAS SRI and grain yield while using the Cropscan system in 2018 and 2019. Dissimilarly, the SVC system in 2020 and 2021 produced moderate correlations across UAS SRI and grain yield. UAS SRI were consistently more heritable, with broad-sense heritability ranging from 0.58 to 0.80. Data standardization and collection windows are important to consider in ensuring reliable data. Furthermore, practical aspects and best practices for these HTP platforms, relative to applied breeding applications, are highlighted and discussed. The findings of this study can be a framework to build upon when considering the implementation of HTP technology in an applied breeding program.

Why it matches plant phenotyping methods複数の地上・UAS型HTPセンサープラットフォームを比較評価し、スペクトル形質の相関・遺伝率・標準化を検証しており、フェノタイピング手法が研究の中心です。

abstractwe evaluated two handheld radiometer platforms, Cropscan MSR16R and Spectra Vista Corp (SVC) HR-1024i, as well as a UAS-based system with a Sentera Quad Multispectral Sensor.

Image-based phenotyping of seed architectural traits and prediction of seed weight using machine learning models in soybean →

Among seed attributes, weight is one of the main factors determining the soybean harvest index. Recently, the focus of soybean breeding has shifted to improving seed size and weight for crop optimization in terms of seed and oil yield. With recent technological advancements, there is an increasing application of imaging sensors that provide simple, real-time, non-destructive, and inexpensive image data for rapid image-based prediction of seed traits in plant breeding programs. The present work is related to digital image analysis of seed traits for the prediction of hundred-seed weight (HSW) in soybean. The image-based seed architectural traits (i-traits) measured were area size (AS), perimeter length (PL), length (L), width (W), length-to-width ratio (LWR), intersection of length and width (IS), seed circularity (CS), and distance between IS and CG (DS). The phenotypic investigation revealed significant genetic variability among 164 soybean genotypes for both i-traits and manually measured seed weight. Seven popular machine learning (ML) algorithms, namely Simple Linear Regression (SLR), Multiple Linear Regression (MLR), Random Forest (RF), Support Vector Regression (SVR), LASSO Regression (LR), Ridge Regression (RR), and Elastic Net Regression (EN), were used to create models that can predict the weight of soybean seeds based on the image-based novel features derived from the Red-Green-Blue (RGB)/visual image. Among the models, random forest and multiple linear regression models that use multiple explanatory variables related to seed size traits (AS, L, W, and DS) were identified as the best models for predicting seed weight with the highest prediction accuracy (coefficient of determination, R 2= 0.98 and 0.94, respectively) and the lowest prediction error, i.e., root mean square error (RMSE) and mean absolute error (MAE). Finally, principal components analysis (PCA) and a hierarchical clustering approach were used to identify IC538070 as a superior genotype with a larger seed size and weight. The identified donors/traits can potentially be used in soybean improvement programs

Why it matches plant phenotyping methodsRGB画像から種子形態形質を抽出し、機械学習で種子重量を予測する画像ベース表現型解析が研究の中心である。

abstractThe present work is related to digital image analysis of seed traits for the prediction of hundred-seed weight (HSW) in soybean.

A novel method for irrigating plants, tracking water use, and imposing water deficits in controlled environments. →

The study of genomic control of drought tolerance in crops requires techniques to impose well defined and consistent levels of drought stress and efficiently measure single-plant water use for hundreds of experimental units over timescales of several months. Traditional gravimetric methods are extremely labor intensive or require expensive technology, and are subject to other errors. This study demonstrates a low-cost, passive, bottom-watered system that is easily scaled for high-throughput phenotyping. The soil water content in the pots is controlled by altering the water table height in an underlying wicking bed via a float valve. The resulting soil moisture profile is then maintained passively as water withdrawn by the plant is replaced by upward movement of water from the wicking bed, which is fed from a reservoir via the float valve. The single-plant water use can be directly measured over time intervals from one to several days by observing the water level in the reservoir. Using this method, four different drought stress levels were induced in pots containing soybean (Glycine max (L.) Merr.), producing four statistically distinct groups for shoot dry weight and seed yield, as well as clear treatment effects for other relevant parameters, including root:shoot dry weight ratio, pod number, cumulative water use, and water use efficiency. This system has a broad range of applications, and should increase feasibility of high-throughput phenotyping efforts for plant drought tolerance traits.

Why it matches plant phenotyping methods高スループット表現型解析のための低コスト灌水・水利用測定システムを開発・実証しており、植物の水利用と乾燥ストレス関連形質の取得が中心的な方法論的貢献である。

abstractThis study demonstrates a low-cost, passive, bottom-watered system that is easily scaled for high-throughput phenotyping.

WheatSpikeNet: an improved wheat spike segmentation model for accurate estimation from field imaging. →

Phenotyping is used in plant breeding to identify genotypes with desirable characteristics, such as drought tolerance, disease resistance, and high-yield potentials. It may also be used to evaluate the effect of environmental circumstances, such as drought, heat, and salt, on plant growth and development. Wheat spike density measure is one of the most important agronomic factors relating to wheat phenotyping. Nonetheless, due to the diversity of wheat field environments, fast and accurate identification for counting wheat spikes remains one of the challenges. This study proposes a meticulously curated and annotated dataset, named as SPIKE-segm, taken from the publicly accessible SPIKE dataset, and an optimal instance segmentation approach named as WheatSpikeNet for segmenting and counting wheat spikes from field imagery. The proposed method is based on the well-known Cascade Mask RCNN architecture with model enhancements and hyperparameter tuning to provide state-of-the-art detection and segmentation performance. A comprehensive ablation analysis incorporating many architectural components of the model was performed to determine the most efficient version. In addition, the model's hyperparameters were fine-tuned by conducting several empirical tests. ResNet50 with Deformable Convolution Network (DCN) as the backbone architecture for feature extraction, Generic RoI Extractor (GRoIE) for RoI pooling, and Side Aware Boundary Localization (SABL) for wheat spike localization comprises the final instance segmentation model. With bbox and mask mean average precision (mAP) scores of 0.9303 and 0.9416, respectively, on the test set, the proposed model achieved superior performance on the challenging SPIKE datasets. Furthermore, in comparison with other existing state-of-the-art methods, the proposed model achieved up to a 0.41% improvement of mAP in spike detection and a significant improvement of 3.46% of mAP in the segmentation tasks that will lead us to an appropriate yield estimation from wheat plants.

Why it matches plant phenotyping methods小麦穂の圃場画像から穂をセグメンテーション・計数する画像解析手法と注釈付きデータセットを開発・評価しており、植物形質取得が研究の中心である。

abstractThis study proposes a meticulously curated and annotated dataset, named as SPIKE-segm, taken from the publicly accessible SPIKE dataset, and an optimal instance segmentation approach named as WheatSpikeNet for segmenting and counting wheat spikes from field imagery.

TrIncNet: a lightweight vision transformer network for identification of plant diseases →

In the agricultural sector, identifying plant diseases at their earliest possible stage of infestation still remains a huge challenge with respect to the maximization of crop production and farmers' income. In recent years, advanced computer vision techniques like Vision Transformers (ViTs) are being successfully applied to identify plant diseases automatically. However, the MLP module in existing ViTs is computationally expensive as well as inefficient in extracting promising features from diseased images. Therefore, this study proposes a comparatively lightweight and improved vision transformer network, also known as "TrIncNet" for plant disease identification. In the proposed network, we introduced a modified encoder architecture a.k.a. Trans-Inception block in which the MLP block of existing ViT was replaced by a custom inception block. Additionally, each Trans-Inception block is surrounded by a skip connection, making it much more resistant to the vanishing gradient problem. The applicability of the proposed network for identifying plant diseases was assessed using two plant disease image datasets viz: PlantVillage dataset and Maize disease dataset (contains in-field images of Maize diseases). The comparative performance analysis on both datasets reported that the proposed TrIncNet network outperformed the state-of-the-art CNN architectures viz: VGG-19, GoogLeNet, ResNet-50, Xception, InceptionV3, and MobileNet. Moreover, the experimental results also showed that the proposed network had achieved 5.38% and 2.87% higher testing accuracy than the existing ViT network on both datasets, respectively. Therefore, the lightweight nature and improved prediction performance make the proposed network suitable for being integrated with IoT devices to assist the stakeholders in identifying plant diseases at the field level.

Why it matches plant phenotyping methods植物の病徴画像から疾病状態を推定する新規画像解析モデルを開発し、複数データセットと既存モデルで性能比較しており、植物フェノタイピング手法が中心である。

abstractthis study proposes a comparatively lightweight and improved vision transformer network, also known as "TrIncNet" for plant disease identification.

Deep reinforcement learning enables adaptive-image augmentation for automated optical inspection of plant rust. →

This study proposes an adaptive image augmentation scheme using deep reinforcement learning (DRL) to improve the performance of a deep learning-based automated optical inspection system. The study addresses the challenge of inconsistency in the performance of single image augmentation methods. It introduces a DRL algorithm, DQN, to select the most suitable augmentation method for each image. The proposed approach extracts geometric and pixel indicators to form states, and uses DeepLab-v3+ model to verify the augmented images and generate rewards. Image augmentation methods are treated as actions, and the DQN algorithm selects the best methods based on the images and segmentation model. The study demonstrates that the proposed framework outperforms any single image augmentation method and achieves better segmentation performance than other semantic segmentation models. The framework has practical implications for developing more accurate and robust automated optical inspection systems, critical for ensuring product quality in various industries. Future research can explore the generalizability and scalability of the proposed framework to other domains and applications. The code for this application is uploaded at https://github.com/lynnkobe/Adaptive-Image-Augmentation.git.

Why it matches plant phenotyping methods植物のさび病画像から病斑をセグメンテーションするための適応的画像拡張法をDRLで開発・評価しており、植物病害状態の画像ベース表現型取得が中心である。

abstractThis study proposes an adaptive image augmentation scheme using deep reinforcement learning (DRL) to improve the performance of a deep learning-based automated optical inspection system.

Allometric equations for estimating peak uprooting force of riparian vegetation. →

Uprooting caused by flood events is a significant disturbance factor that affects the establishment, growth, and mortality of riparian vegetation. If the hydraulic drag force acting on riparian plants exceeds the peak uprooting force originate from their below-ground portion, it may result in the uprooting of these plants. Despite previous studies have documented and investigated the uprooting processes and factors influencing the peak uprooting force of plants, most of these studies have focused on how the root morphological traits of tree and shrub seedlings affect peak uprooting force or mainly collected data in indoor experiments, which may limit the extrapolation of the results to natural environments. To address these limitations, we assume that the peak uprooting force can be estimated by the morphological traits of the above-ground portion of the vegetation. In this study, we conducted in-situ vertical uprooting tests on three locally dominant species: Conyza canadensis , Daucus carota , and Leonurus sibiricus , in a typical riverine environment. The three species were found to have the highest abundance based on the outcomes of the quadrat method. We measured the peak uprooting force, plant height, stem basal diameter, shoot and root wet biomass, and shoot and root dry biomass of each plant and compared them between species. Furthermore, we quantified the influence of morphology on peak uprooting force. Our results showed significant differences in morphological traits and peak uprooting force among the three species. We found a significant positive correlation between peak uprooting force and the morphological traits of the three species. The peak uprooting force increases with plant size following a power law function which is analogous to allometric equations. The allometric equation provided a convenient and non-destructive method to estimate the peak uprooting force based on the above-ground morphological traits of the plants, which may help to overcome the limitations of measuring root morphological traits.

Why it matches plant phenotyping methods植物の地上部形態から地下部に由来する最大引抜抵抗力を推定する非破壊的なアロメトリック手法を開発・提示しており、形質取得・推定が研究の中心である。

abstractThe allometric equation provided a convenient and non-destructive method to estimate the peak uprooting force based on the above-ground morphological traits of the plants

Row selection in remote sensing from four-row plots of maize and sorghum based on repeatability and predictive modeling. →

Remote sensing enables the rapid assessment of many traits that provide valuable information to plant breeders throughout the growing season to improve genetic gain. These traits are often extracted from remote sensing data on a row segment (rows within a plot) basis enabling the quantitative assessment of any row-wise subset of plants in a plot, rather than a few individual representative plants, as is commonly done in field-based phenotyping. Nevertheless, which rows to include in analysis is still a matter of debate. The objective of this experiment was to evaluate row selection and plot trimming in field trials conducted using four-row plots with remote sensing traits extracted from RGB (red-green-blue), LiDAR (light detection and ranging), and VNIR (visible near infrared) hyperspectral data. Uncrewed aerial vehicle flights were conducted throughout the growing seasons of 2018 to 2021 with data collected on three years of a sorghum experiment and two years of a maize experiment. Traits were extracted from each plot based on all four row segments (RS) (RS1234), inner rows (RS23), outer rows (RS14), and individual rows (RS1, RS2, RS3, and RS4). Plot end trimming of 40 cm was an additional factor tested. Repeatability and predictive modeling of end-season yield were used to evaluate performance of these methodologies. Plot trimming was never shown to result in significantly different outcomes from non-trimmed plots. Significant differences were often observed based on differences in row selection. Plots with more row segments were often favorable for increasing repeatability, and excluding outer rows improved predictive modeling. These results support long-standing principles of experimental design in agronomy and should be considered in breeding programs that incorporate remote sensing.

Why it matches plant phenotyping methodsRGB・LiDAR・VNIRリモートセンシングによる作物形質抽出について、行選択とプロットトリミングを反復性・収量予測で評価しており、フェノタイピング手法の技術評価が中心である。

abstractThe objective of this experiment was to evaluate row selection and plot trimming in field trials conducted using four-row plots with remote sensing traits extracted from RGB (red-green-blue), LiDAR (light detection and ranging), and VNIR (visible near infrared) hyperspectral data.

CropQuant-Air: an AI-powered system to enable phenotypic analysis of yield- and performance-related traits using wheat canopy imagery collected by low-cost drones. →

As one of the most consumed stable foods around the world, wheat plays a crucial role in ensuring global food security. The ability to quantify key yield components under complex field conditions can help breeders and researchers assess wheat’s yield performance effectively. Nevertheless, it is still challenging to conduct large-scale phenotyping to analyse canopy-level wheat spikes and relevant performance traits, in the field and in an automated manner. Here, we present CropQuant-Air, an AI-powered software system that combines state-of-the-art deep learning (DL) models and image processing algorithms to enable the detection of wheat spikes and phenotypic analysis using wheat canopy images acquired by low-cost drones. The system includes the YOLACT-Plot model for plot segmentation, an optimised YOLOv7 model for quantifying the spike number per m2(SNpM2) trait, and performance-related trait analysis using spectral and texture features at the canopy level. Besides using our labelled dataset for model training, we also employed the Global Wheat Head Detection dataset to incorporate varietal features into the DL models, facilitating us to perform reliable yield-based analysis from hundreds of varieties selected from main wheat production regions in China. Finally, we employed the SNpM2and performance traits to develop a yield classification model using the Extreme Gradient Boosting (XGBoost) ensemble and obtained significant positive correlations between the computational analysis results and manual scoring, indicating the reliability of CropQuant-Air. To ensure that our work could reach wider researchers, we created a graphical user interface for CropQuant-Air, so that non-expert users could readily use our work. We believe that our work represents valuable advances in yield-based field phenotyping and phenotypic analysis, providing useful and reliable toolkits to enable breeders, researchers, growers, and farmers to assess crop-yield performance in a cost-effective approach.

Why it matches plant phenotyping methodsドローン画像からコムギ穂数や収量関連形質を抽出・検証するAIソフトウェア／表現型解析システムが研究の中心であり、植物フェノタイピング手法に該当する。

abstractwe present CropQuant-Air, an AI-powered software system that combines state-of-the-art deep learning (DL) models and image processing algorithms to enable the detection of wheat spikes and phenotypic analysis using wheat canopy images acquired by low-cost drones.

Sainfoin (Onobrychis spp.) crop ontology: supporting germplasm characterization and international research collaborations →

Sainfoin ( Onobrychis spp.) is a perennial forage legume that is also attracting attention as a perennial pulse with potential for human consumption. The dual use of sainfoin underpins diverse research and breeding programs focused on improving sainfoin lines for forage and pulses, which is driving the generation of complex datasets describing high dimensional phenotypes in the post-omics era. To ensure that multiple user groups, for example, breeders selecting for forage and those selecting for edible seed, can utilize these rich datasets, it is necessary to develop common ontologies and accessible ontology platforms. One such platform, Crop Ontology, was created in 2008 by the Consortium of International Agricultural Research Centers (CGIAR) to host crop-specific trait ontologies that support standardized plant breeding databases. In the present study, we describe the sainfoin crop ontology (CO). An in-depth literature review was performed to develop a comprehensive list of traits measured and reported in sainfoin. Because the same traits can be measured in different ways, ultimately, a set of 98 variables (variable = plant trait + method of measurement + scale of measurement) used to describe variation in sainfoin were identified. Variables were formatted and standardized based on guidelines provided here for inclusion in the sainfoin CO. The 98 variables contained a total of 82 traits from four trait classes of which 24 were agronomic, 31 were morphological, 19 were seed and forage quality related, and 8 were phenological. In addition to the developed variables, we have provided a roadmap for developing and submission of new traits to the sainfoin CO.

Why it matches plant phenotyping methods植物形質の測定方法と尺度を含む変数を標準化し、育種データベースで再利用可能な作物オントロジーを開発した研究であり、形質データ基盤が中心です。

abstractwe describe the sainfoin crop ontology (CO)

Digital whole-community phenotyping: tracking morphological and physiological responses of plant communities to environmental changes in the field →

Plant traits are informative for ecosystem functions and processes and help to derive general rules and predictions about responses to environmental gradients, global change and perturbations. Ecological field studies often use 'low-throughput' methods to assess plant phenotypes and integrate species-specific traits to community-wide indices. In contrast, agricultural greenhouse or lab-based studies often employ 'high-throughput phenotyping' to assess plant individuals tracking their growth or fertilizer and water demand. In ecological field studies, remote sensing makes use of freely movable devices like satellites or unmanned aerial vehicles (UAVs) which provide large-scale spatial and temporal data. Adopting such methods for community ecology on a smaller scale may provide novel insights on the phenotypic properties of plant communities and fill the gap between traditional field measurements and airborne remote sensing. However, the trade-off between spatial resolution, temporal resolution and scope of the respective study requires highly specific setups so that the measurements fit the scientific question. We introduce small-scale, high-resolution digital automated phenotyping as a novel source of quantitative trait data in ecological field studies that provides complementary multi-faceted data of plant communities. We customized an automated plant phenotyping system for its mobile application in the field for 'digital whole-community phenotyping' (DWCP), capturing the 3-dimensional structure and multispectral information of plant communities. We demonstrated the potential of DWCP by recording plant community responses to experimental land-use treatments over two years. DWCP captured changes in morphological and physiological community properties in response to mowing and fertilizer treatments and thus reliably informed about changes in land-use. In contrast, manually measured community-weighted mean traits and species composition remained largely unaffected and were not informative about these treatments. DWCP proved to be an efficient method for characterizing plant communities, complements other methods in trait-based ecology, provides indicators of ecosystem states, and may help to forecast tipping points in plant communities often associated with irreversible changes in ecosystems.

Why it matches plant phenotyping methods植物群集の3次元構造とマルチスペクトル情報を取得する自動フェノタイピングシステムをフィールド用に改変・実証しており、植物表現型取得法が研究の中心である。

abstractWe introduce small-scale, high-resolution digital automated phenotyping as a novel source of quantitative trait data in ecological field studies

LightMixer: A novel lightweight convolutional neural network for tomato disease detection. →

Tomatoes are among the very important crops grown worldwide. However, tomato diseases can harm the health of tomato plants during growth and reduce tomato yields over large areas. The development of computer vision technology offers the prospect of solving this problem. However, traditional deep learning algorithms require a high computational cost and several parameters. Therefore, a lightweight tomato leaf disease identification model called LightMixer was designed in this study. The LightMixer model comprises a depth convolution with a Phish module and a light residual module. Depth convolution with the Phish module represents a lightweight convolution module designed to splice nonlinear activation functions with depth convolution as the backbone; it also focuses on lightweight convolutional feature extraction to facilitate deep feature fusion. The light residual module was built based on lightweight residual blocks to accelerate the computational efficiency of the entire network architecture and reduce the information loss of disease features. Experimental results show that the proposed LightMixer model achieved 99.3% accuracy on public datasets while requiring only 1.5 M parameters, an improvement over other classical convolutional neural network and lightweight models, and can be used for automatic tomato leaf disease identification on mobile devices.

Why it matches plant phenotyping methodsトマト葉の病徴を画像から識別する軽量CNNを開発し、精度・計算量を比較評価しており、植物病害状態の取得・推定手法が中心である。

abstracta lightweight tomato leaf disease identification model called LightMixer was designed in this study.

Self-supervised maize kernel classification and segmentation for embryo identification. →

Introduction Computer vision and deep learning (DL) techniques have succeeded in a wide range of diverse fields. Recently, these techniques have been successfully deployed in plant science applications to address food security, productivity, and environmental sustainability problems for a growing global population. However, training these DL models often necessitates the large-scale manual annotation of data which frequently becomes a tedious and time-and-resource- intensive process. Recent advances in self-supervised learning (SSL) methods have proven instrumental in overcoming these obstacles, using purely unlabeled datasets to pre-train DL models. Methods Here, we implement the popular self-supervised contrastive learning methods of NNCLR Nearest neighbor Contrastive Learning of visual Representations) and SimCLR (Simple framework for Contrastive Learning of visual Representations) for the classification of spatial orientation and segmentation of embryos of maize kernels. Maize kernels are imaged using a commercial high-throughput imaging system. This image data is often used in multiple downstream applications across both production and breeding applications, for instance, sorting for oil content based on segmenting and quantifying the scutellum's size and for classifying haploid and diploid kernels. Results and discussion We show that in both classification and segmentation problems, SSL techniques outperform their purely supervised transfer learning-based counterparts and are significantly more annotation efficient. Additionally, we show that a single SSL pre-trained model can be efficiently finetuned for both classification and segmentation, indicating good transferability across multiple downstream applications. Segmentation models with SSL-pretrained backbones produce DICE similarity coefficients of 0.81, higher than the 0.78 and 0.73 of those with ImageNet-pretrained and randomly initialized backbones, respectively. We observe that finetuning classification and segmentation models on as little as 1% annotation produces competitive results. These results show SSL provides a meaningful step forward in data efficiency with agricultural deep learning and computer vision.

Why it matches plant phenotyping methodsトウモロコシ穀粒画像から胚の向き分類と分割を行う自己教師あり画像解析手法を開発・評価しており、植物器官の表現型抽出が研究の中心である。

abstractwe implement the popular self-supervised contrastive learning methods of NNCLR Nearest neighbor Contrastive Learning of visual Representations) and SimCLR (Simple framework for Contrastive Learning of visual Representations) for the classification of spatial orientation and segmentation of embryos of maize kernels.

BAF-Net: Bidirectional attention fusion network via CNN and transformers for the pepper leaf segmentation. →

The segmentation of pepper leaves from pepper images is of great significance for the accurate control of pepper leaf diseases. To address the issue, we propose a bidirectional attention fusion network combing the convolution neural network (CNN) and Swin Transformer, called BAF-Net, to segment the pepper leaf image. Specially, BAF-Net first uses a multi-scale fusion feature (MSFF) branch to extract the long-range dependencies by constructing the cascaded Swin Transformer-based and CNN-based block, which is based on the U-shape architecture. Then, it uses a full-scale feature fusion (FSFF) branch to enhance the boundary information and attain the detailed information. Finally, an adaptive bidirectional attention module is designed to bridge the relation of the MSFF and FSFF features. The results on four pepper leaf datasets demonstrated that our model obtains F1 scores of 96.75%, 91.10%, 97.34% and 94.42%, and IoU of 95.68%, 86.76%, 96.12% and 91.44%, respectively. Compared to the state-of-the-art models, the proposed model achieves better segmentation performance. The code will be available at the website: https://github.com/fangchj2002/BAF-Net.

Why it matches plant phenotyping methodsペッパー葉画像から葉領域を抽出する画像解析手法を開発・検証しており、植物器官の形態状態取得が研究の中心である。

abstractwe propose a bidirectional attention fusion network combing the convolution neural network (CNN) and Swin Transformer, called BAF-Net, to segment the pepper leaf image.

PhytoOracle: Scalable, modular phenomics data processing pipelines →

As phenomics data volume and dimensionality increase due to advancements in sensor technology, there is an urgent need to develop and implement scalable data processing pipelines. Current phenomics data processing pipelines lack modularity, extensibility, and processing distribution across sensor modalities and phenotyping platforms. To address these challenges, we developed PhytoOracle (PO), a suite of modular, scalable pipelines for processing large volumes of field phenomics RGB, thermal, PSII chlorophyll fluorescence 2D images, and 3D point clouds. PhytoOracle aims to ( i ) improve data processing efficiency; ( ii ) provide an extensible, reproducible computing framework; and ( iii ) enable data fusion of multi-modal phenomics data. PhytoOracle integrates open-source distributed computing frameworks for parallel processing on high-performance computing, cloud, and local computing environments. Each pipeline component is available as a standalone container, providing transferability, extensibility, and reproducibility. The PO pipeline extracts and associates individual plant traits across sensor modalities and collection time points, representing a unique multi-system approach to addressing the genotype-phenotype gap. To date, PO supports lettuce and sorghum phenotypic trait extraction, with a goal of widening the range of supported species in the future. At the maximum number of cores tested in this study (1,024 cores), PO processing times were: 235 minutes for 9,270 RGB images (140.7 GB), 235 minutes for 9,270 thermal images (5.4 GB), and 13 minutes for 39,678 PSII images (86.2 GB). These processing times represent end-to-end processing, from raw data to fully processed numerical phenotypic trait data. Repeatability values of 0.39-0.95 (bounding area), 0.81-0.95 (axis-aligned bounding volume), 0.79-0.94 (oriented bounding volume), 0.83-0.95 (plant height), and 0.81-0.95 (number of points) were observed in Field Scanalyzer data. We also show the ability of PO to process drone data with a repeatability of 0.55-0.95 (bounding area).

Why it matches plant phenotyping methods植物フェノミクスのマルチモーダル画像・点群から形質を抽出する、スケーラブルで再現可能な処理パイプラインの開発と反復性評価が中心である。

abstractwe developed PhytoOracle (PO), a suite of modular, scalable pipelines for processing large volumes of field phenomics RGB, thermal, PSII chlorophyll fluorescence 2D images, and 3D point clouds.

Detection of peanut seed vigor based on hyperspectral imaging and chemometrics. →

Rapid nondestructive testing of peanut seed vigor is of great significance in current research. Before seeds are sown, effective screening of high-quality seeds for planting is crucial to improve the quality of crop yield, and seed vitality is one of the important indicators to evaluate seed quality, which can represent the potential ability of seeds to germinate quickly and whole and grow into normal seedlings or plants. Meanwhile, the advantage of nondestructive testing technology is that the seeds themselves will not be damaged. In this study, hyperspectral technology and superoxide dismutase activity were used to detect peanut seed vigor. To investigate peanut seed vigor and predict superoxide dismutase activity, spectral characteristics of peanut seeds in the wavelength range of 400-1000 nm were analyzed. The spectral data are processed by a variety of hot spot algorithms. Spectral data were preprocessed with Savitzky-Golay (SG), multivariate scatter correction (MSC), and median filtering (MF), which can effectively to reduce the effects of baseline drift and tilt. CatBoost and Gradient Boosted Decision Tree were used for feature band extraction, the top five weights of the characteristic bands of peanut seed vigor classification are 425.48nm, 930.8nm, 965.32nm, 984.0nm, and 994.7nm. XGBoost, LightGBM, Support Vector Machine and Random Forest were used for modeling of seed vitality classification. XGBoost and partial least squares regression were used to establish superoxide dismutase activity value regression model. The results indicated that MF-CatBoost-LightGBM was the best model for peanut seed vigor classification, and the accuracy result was 90.83%. MSC-CatBoost-PLSR was the optimal regression model of superoxide dismutase activity value. The results show that the R 2 was 0.9787 and the RMSE value was 0.0566. The results suggested that hyperspectral technology could correlate the external manifestation of effective peanut seed vigor.

Why it matches plant phenotyping methods落花生種子の活力という植物形質をハイパースペクトル画像と機械学習で非破壊推定する手法が研究の中心であり、分類・回帰性能も評価している。

abstractIn this study, hyperspectral technology and superoxide dismutase activity were used to detect peanut seed vigor.

EADD-YOLO: An efficient and accurate disease detector for apple leaf using improved lightweight YOLOv5. →

Introduction Current detection methods for apple leaf diseases still suffer some challenges, such as the high number of parameters, low detection speed and poor detection performance for small dense spots, which limit the practical applications in agriculture. Therefore, an efficient and accurate model for apple leaf disease detection based on YOLOv5 is proposed and named EADD-YOLO. Methods In the EADD-YOLO, the lightweight shufflenet inverted residual module is utilized to reconstruct the backbone network, and an efficient feature learning module designed through depthwise convolution is proposed and introduced to the neck network. The aim is to reduce the number of parameters and floating point of operations (FLOPs) during feature extraction and feature fusion, thus increasing the operational efficiency of the network with less impact on detection performance. In addition, the coordinate attention module is embedded into the critical locations of the network to select the critical spot information and suppress useless information, which is to enhance the detection accuracy of diseases with various sizes from different scenes. Furthermore, the SIoU loss replaces CIoU loss as the bounding box regression loss function to improve the accuracy of prediction box localization. Results The experimental results indicate that the proposed method can achieve the detection performance of 95.5% on the mean average precision and a speed of 625 frames per second (FPS) on the apple leaf disease dataset (ALDD). Compared to the latest research method on the ALDD, the detection accuracy and speed of the proposed method were improved by 12.3% and 596 FPS, respectively. In addition, the parameter quantity and FLOPs of the proposed method were much less than other relevant popular algorithms. Discussion In summary, the proposed method not only has a satisfactory detection effect, but also has fewer parameters and high calculation efficiency compared with the existing approaches. Therefore, the proposed method provides a high-performance solution for the early diagnosis of apple leaf disease and can be applied in agricultural robots. The code repository is open-sourced at https://github.com/AWANWY/EADD-YOLO.

Why it matches plant phenotyping methodsリンゴ葉の病斑を画像から検出するYOLOベース手法の開発・性能評価が研究の中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法として適格。

abstractTherefore, an efficient and accurate model for apple leaf disease detection based on YOLOv5 is proposed and named EADD-YOLO.

EmergeNet: A novel deep-learning based ensemble segmentation model for emergence timing detection of coleoptile. →

The emergence timing of a plant, i.e., the time at which the plant is first visible from the surface of the soil, is an important phenotypic event and is an indicator of the successful establishment and growth of a plant. The paper introduces a novel deep-learning based model called EmergeNet with a customized loss function that adapts to plant growth for coleoptile (a rigid plant tissue that encloses the first leaves of a seedling) emergence timing detection. It can also track its growth from a time-lapse sequence of images with cluttered backgrounds and extreme variations in illumination. EmergeNet is a novel ensemble segmentation model that integrates three different but promising networks, namely, SEResNet, InceptionV3, and VGG19, in the encoder part of its base model, which is the UNet model. EmergeNet can correctly detect the coleoptile at its first emergence when it is tiny and therefore barely visible on the soil surface. The performance of EmergeNet is evaluated using a benchmark dataset called the University of Nebraska-Lincoln Maize Emergence Dataset (UNL-MED). It contains top-view time-lapse images of maize coleoptiles starting before the occurrence of their emergence and continuing until they are about one inch tall. EmergeNet detects the emergence timing with 100% accuracy compared with human-annotated ground-truth. Furthermore, it significantly outperforms UNet by generating very high-quality segmented masks of the coleoptiles in both natural light and dark environmental conditions.

Why it matches plant phenotyping methodsコレオプタイルの出芽時期と成長を画像から抽出する深層学習セグメンテーション手法を開発し、ベンチマークデータセットで性能検証しているため、植物フェノタイピング手法が中心である。

abstractThe performance of EmergeNet is evaluated using a benchmark dataset called the University of Nebraska-Lincoln Maize Emergence Dataset (UNL-MED).

Automatic counting of rapeseed inflorescences using deep learning method and UAV RGB imagery. →

Flowering is a crucial developing stage for rapeseed ( Brassica napus L.) plants. Flowers develop on the main and branch inflorescences of rapeseed plants and then grow into siliques. The seed yield of rapeseed heavily depends on the total flower numbers per area throughout the whole flowering period. The number of rapeseed inflorescences can reflect the richness of rapeseed flowers and provide useful information for yield prediction. To count rapeseed inflorescences automatically, we transferred the counting problem to a detection task. Then, we developed a low-cost approach for counting rapeseed inflorescences using YOLOv5 with the Convolutional Block Attention Module (CBAM) based on unmanned aerial vehicle (UAV) Red-Green-Blue (RGB) imagery. Moreover, we constructed a Rapeseed Inflorescence Benchmark (RIB) to verify the effectiveness of our model. The RIB dataset captured by DJI Phantom 4 Pro V2.0, including 165 plot images and 60,000 manual labels, is to be released. Experimental results showed that indicators R 2 for counting and the mean Average Precision (mAP) for location were over 0.96 and 92%, respectively. Compared with Faster R-CNN, YOLOv4, CenterNet, and TasselNetV2+, the proposed method achieved state-of-the-art counting performance on RIB and had advantages in location accuracy. The counting results revealed a quantitative dynamic change in the number of rapeseed inflorescences in the time dimension. Furthermore, a significant positive correlation between the actual crop yield and the automatically obtained rapeseed inflorescence total number on a field plot level was identified. Thus, a set of UAV- assisted methods for better determination of the flower richness was developed, which can greatly support the breeding of high-yield rapeseed varieties.

Why it matches plant phenotyping methodsUAV RGB画像と深層学習を用いてナタネの花序数を自動計数する手法を開発し、ベンチマークデータセットで検証しているため、植物表現型取得が中心である。

abstractwe developed a low-cost approach for counting rapeseed inflorescences using YOLOv5 with the Convolutional Block Attention Module (CBAM) based on unmanned aerial vehicle (UAV) Red-Green-Blue (RGB) imagery.

Non-coding deep learning models for tomato biotic and abiotic stress classification using microscopic images. →

Plant disease classification is quite complex and, in most cases, requires trained plant pathologists and sophisticated labs to accurately determine the cause. Our group for the first time used microscopic images (×30) of tomato plant diseases, for which representative plant samples were diagnostically validated to classify disease symptoms using non-coding deep learning platforms (NCDL). The mean F1 scores (SD) of the NCDL platforms were 98.5 (1.6) for Amazon Rekognition Custom Label, 93.9 (2.5) for Clarifai, 91.6 (3.9) for Teachable Machine, 95.0 (1.9) for Google AutoML Vision, and 97.5 (2.7) for Microsoft Azure Custom Vision. The accuracy of the NCDL platform for Amazon Rekognition Custom Label was 99.8% (0.2), for Clarifai 98.7% (0.5), for Teachable Machine 98.3% (0.4), for Google AutoML Vision 98.9% (0.6), and for Apple CreateML 87.3 (4.3). Upon external validation, the model's accuracy of the tested NCDL platforms dropped no more than 7%. The potential future use for these models includes the development of mobile- and web-based applications for the classification of plant diseases and integration with a disease management advisory system. The NCDL models also have the potential to improve the early triage of symptomatic plant samples into classes that may save time in diagnostic lab sample processing.

Why it matches plant phenotyping methodsトマト葉の顕微鏡画像から病徴を分類する深層学習モデルを開発・比較し、外部検証まで実施しており、植物病害状態の表現型取得が中心である。

abstractUpon external validation, the model's accuracy of the tested NCDL platforms dropped no more than 7%.

Visual question answering model for fruit tree disease decision-making based on multimodal deep learning. →

Visual Question Answering (VQA) about diseases is an essential feature of intelligent management in smart agriculture. Currently, research on fruit tree diseases using deep learning mainly uses single-source data information, such as visible images or spectral data, yielding classification and identification results that cannot be directly used in practical agricultural decision-making. In this study, a VQA model for fruit tree diseases based on multimodal feature fusion was designed. Fusing images and Q&A knowledge of disease management, the model obtains the decision-making answer by querying questions about fruit tree disease images to find relevant disease image regions. The main contributions of this study were as follows: (1) a multimodal bilinear factorized pooling model using Tucker decomposition was proposed to fuse the image features with question features: (2) a deep modular co-attention architecture was explored to simultaneously learn the image and question attention to obtain richer graphical features and interactivity. The experiments showed that the proposed unified model combining the bilinear model and co-attentive learning in a new network architecture obtained 86.36% accuracy in decision-making under the condition of limited data (8,450 images and 4,560k Q&A pairs of data), outperforming existing multimodal methods. The data augmentation is adopted on the training set to avoid overfitting. Ten runs of 10-fold cross-validation are used to report the unbiased performance. The proposed multimodal fusion model achieved friendly interaction and fine-grained identification and decision-making performance. Thus, the model can be widely deployed in intelligent agriculture.

Why it matches plant phenotyping methods果樹病害画像から病害状態を推定するマルチモーダルVQAモデルの開発と性能検証が研究の中心であり、植物病害表現型の画像ベース推定に該当する。

abstractIn this study, a VQA model for fruit tree diseases based on multimodal feature fusion was designed.

Research on cassava disease classification using the multi-scale fusion model based on EfficientNet and attention mechanism. →

Cassava disease is one of the leading causes to the serious decline of cassava yield. Because it is difficult to identify the characteristics of cassava disease, if not professional cassava growers, it will be prone to misjudgment. In order to strengthen the judgment of cassava diseases, the identification characteristics of cassava diseases such as different color of cassava leaf disease spots, abnormal leaf shape and disease spot area were studied. In this paper, deep convolutional neural network was used to classify cassava leaf diseases, and image classification technology was used to recognize and classify cassava leaf diseases. A lightweight module Multi-scale fusion model (MSFM) based on attention mechanism was proposed to extract disease features of cassava leaves to enhance the classification of disease features. The resulting feature map contained key disease identification information. The study used 22,000 cassava disease leaf images as a data set, including four different cassava leaf disease categories and healthy cassava leaves. The experimental results show that the cassava leaf disease classification model based on multi-scale fusion Convolutional Neural Network (CNN) improves EfficientNet compared with the original model, with the average recognition rate increased by nearly 4% and the average recognition rate up to 88.1%. It provides theoretical support and practical tools for the recognition and early diagnosis of plant disease leaves.

Why it matches plant phenotyping methodsキャッサバ葉の病斑・形状・面積という植物状態を画像から分類する深層学習手法を提案・評価しており、病害表現型の取得・推定が中心である。

abstractIn this paper, deep convolutional neural network was used to classify cassava leaf diseases, and image classification technology was used to recognize and classify cassava leaf diseases.

Root system architecture and environmental flux analysis in mature crops using 3D root mesocosms →

Current methods of root sampling typically only obtain small or incomplete sections of root systems and do not capture their true complexity. To facilitate the visualization and analysis of full-sized plant root systems in 3-dimensions, we developed customized mesocosm growth containers. While highly scalable, the design presented here uses an internal volume of 45 ft 3 (1.27 m 3 ), suitable for large crop and bioenergy grass root systems to grow largely unconstrained. Furthermore, they allow for the excavation and preservation of 3-dimensional root system architecture (RSA), and facilitate the collection of time-resolved subterranean environmental data. Sensor arrays monitoring matric potential, temperature and CO 2 levels are buried in a grid formation at various depths to assess environmental fluxes at regular intervals. Methods of 3D data visualization of fluxes were developed to allow for comparison with root system architectural traits. Following harvest, the recovered root system can be digitally reconstructed in 3D through photogrammetry, which is an inexpensive method requiring only an appropriate studio space and a digital camera. We developed a pipeline to extract features from the 3D point clouds, or from derived skeletons that include point cloud voxel number as a proxy for biomass, total root system length, volume, depth, convex hull volume and solidity as a function of depth. Ground-truthing these features with biomass measurements from manually dissected root systems showed a high correlation. We evaluated switchgrass, maize, and sorghum root systems to highlight the capability for species wide comparisons. We focused on two switchgrass ecotypes, upland (VS16) and lowland (WBC3), in identical environments to demonstrate widely different root system architectures that may be indicative of core differences in their rhizoeconomic foraging strategies. Finally, we imposed a strong physiological water stress and manipulated the growth medium to demonstrate whole root system plasticity in response to environmental stimuli. Hence, these new "3D Root Mesocosms" and accompanying computational analysis provides a new paradigm for study of mature crop systems and the environmental fluxes that shape them.

Why it matches plant phenotyping methods3Dルートメソコスム、フォトグラメトリ、点群解析による根系形態形質の取得・検証が研究の中心であり、植物フェノタイピング手法に該当する。

abstractTo facilitate the visualization and analysis of full-sized plant root systems in 3-dimensions, we developed customized mesocosm growth containers.

Prediction of heading date, culm length, and biomass from canopy-height-related parameters derived from time-series UAV observations of rice. →

Unmanned aerial vehicles (UAVs) are powerful tools for monitoring crops for high-throughput phenotyping. Time-series aerial photography of fields can record the whole process of crop growth. Canopy height (CH), which is vertical plant growth, has been used as an indicator for the evaluation of lodging tolerance and the prediction of biomass and yield. However, there have been few attempts to use UAV-derived time-series CH data for field testing of crop lines. Here we provide a novel framework for trait prediction using CH data in rice. We generated UAV-based digital surface models of crops to extract CH data of 30 Japanese rice cultivars in 2019, 2020, and 2021. CH-related parameters were calculated in a non-linear time-series model as an S-shaped plant growth curve. The maximum saturation CH value was the most important predictor for culm length. The time point at the maximum CH contributed to the prediction of days to heading, and was able to predict stem and leaf weight and aboveground weight, possibly reflecting the association of biomass with duration of vegetative growth. These results indicate that the CH-related parameters acquired by UAV can be useful as predictors of traits typically measured by hand.

Why it matches plant phenotyping methodsUAV画像から作物のキャノピー高を抽出し、時系列モデルで生育・形質を予測する枠組みが研究の中心であり、実質的な植物フェノタイピング手法の開発・適用に該当する。

abstractHere we provide a novel framework for trait prediction using CH data in rice.

Using phenomics to identify and integrate traits of interest for better-performing common beans: A validation study on an interspecific hybrid and its Acutifolii parents →

Introduction Evaluations of interspecific hybrids are limited, as classical genebank accession descriptors are semi-subjective, have qualitative traits and show complications when evaluating intermediate accessions. However, descriptors can be quantified using recognized phenomic traits. This digitalization can identify phenomic traits which correspond to the percentage of parental descriptors remaining expressed/visible/measurable in the particular interspecific hybrid. In this study, a line of P. vulgaris , P. acutifolius and P. parvifolius accessions and their crosses were sown in the mesh house according to CIAT seed regeneration procedures. Methodology Three accessions and one derived breeding line originating from their interspecific crosses were characterized and classified by selected phenomic descriptors using multivariate and machine learning techniques. The phenomic proportions of the interspecific hybrid (line INB 47) with respect to its three parent accessions were determined using a random forest and a respective confusion matrix. Results The seed and pod morphometric traits, physiological behavior and yield performance were evaluated. In the classification of the accession, the phenomic descriptors with highest prediction force were Fm', Fo', Fs', LTD, Chl, seed area, seed height, seed Major, seed MinFeret, seed Minor, pod AR, pod Feret, pod round, pod solidity, pod area, pod major, pod seed weight and pod weight. Physiological traits measured in the interspecific hybrid present 2.2% similarity with the P. acutifolius and 1% with the P. parvifolius accessions. In addition, in seed morphometric characteristics, the hybrid showed 4.5% similarity with the P. acutifolius accession. Conclusions Here we were able to determine the phenomic proportions of individual parents in their interspecific hybrid accession. After some careful generalization the methodology can be used to: i) verify trait-of-interest transfer from P. acutifolius and P. parvifolius accessions into their hybrids; ii) confirm selected traits as "phenomic markers" which would allow conserving desired physiological traits of exotic parental accessions, without losing key seed characteristics from elite common bean accessions; and iii) propose a quantitative tool that helps genebank curators and breeders to make better-informed decisions based on quantitative analysis.

Why it matches plant phenotyping methodsインタースペシフィック雑種の形質を定量化・分類し、ランダムフォレストと混同行列で親由来のフェノミック形質割合を検証する方法論が研究の中心である。

abstractThis digitalization can identify phenomic traits which correspond to the percentage of parental descriptors remaining expressed/visible/measurable in the particular interspecific hybrid.

Optimization of a static headspace GC-MS method and its application in metabolic fingerprinting of the leaf volatiles of 42 citrus cultivars. →

Citrus leaves, which are a rich source of plant volatiles, have the beneficial attributes of rapid growth, large biomass, and availability throughout the year. Establishing the leaf volatile profiles of different citrus genotypes would make a valuable contribution to citrus species identification and chemotaxonomic studies. In this study, we developed an efficient and convenient static headspace (HS) sampling technique combined with gas chromatography-mass spectrometry (GC-MS) analysis and optimized the extraction conditions (a 15-min incubation at 100 ˚C without the addition of salt). Using a large set of 42 citrus cultivars, we validated the applicability of the optimized HS-GC-MS system in determining leaf volatile profiles. A total of 83 volatile metabolites, including monoterpene hydrocarbons, alcohols, sesquiterpene hydrocarbons, aldehydes, monoterpenoids, esters, and ketones were identified and quantified. Multivariate statistical analysis and hierarchical clustering revealed that mandarin ( Citrus reticulata Blanco) and orange ( Citrus sinensis L. Osbeck) groups exhibited notably differential volatile profiles, and that the mandarin group cultivars were characterized by the complex volatile profiles, thereby indicating the complex nature and diversity of these mandarin cultivars. We also identified those volatile compounds deemed to be the most useful in discriminating amongst citrus cultivars. This method developed in this study provides a rapid, simple, and reliable approach for the extraction and identification of citrus leaf volatile organic compound, and based on this methodology, we propose a leaf volatile profile-based classification model for citrus.

Why it matches plant phenotyping methods葉の揮発性化合物プロファイルを取得・識別する分析法の最適化と、42品種での適用性検証が研究の中心であり、植物器官の化学的表現型を測定する再利用可能な方法を提示している。

abstractwe developed an efficient and convenient static headspace (HS) sampling technique combined with gas chromatography-mass spectrometry (GC-MS) analysis and optimized the extraction conditions

Data-centric annotation analysis for plant disease detection: Strategy, consistency, and performance. →

Object detection models have become the current tool of choice for plant disease detection in precision agriculture. Most existing research improved the performance by ameliorating networks and optimizing the loss function. However, because of the vast influence of data annotation quality and the cost of annotation, the data-centric part of a project also needs more investigation. We should further consider the relationship between data annotation strategies, annotation quality, and the model's performance. In this paper, a systematic strategy with four annotation strategies for plant disease detection is proposed: local, semi-global, global, and symptom-adaptive annotation. Labels with different annotation strategies will result in distinct models' performance, and their contrasts are remarkable. An interpretability study of the annotation strategy is conducted by using class activation maps. In addition, we define five types of inconsistencies in the annotation process and investigate the severity of the impact of inconsistent labels on model's performance. Finally, we discuss the problem of label inconsistency during data augmentation. Overall, this data-centric quantitative analysis helps us to understand the significance of annotation strategies, which provides practitioners a way to obtain higher performance and reduce annotation costs on plant disease detection. Our work encourages researchers to pay more attention to annotation consistency and the essential issues of annotation strategy. The code will be released at: https://github.com/JiuqingDong/PlantDiseaseDetection_Yolov5 .

Why it matches plant phenotyping methods植物病害の画像検出におけるアノテーション戦略・一貫性とモデル性能を体系的に評価する手法研究であり、植物の病徴・病害状態の抽出方法が中心的である。

abstracta systematic strategy with four annotation strategies for plant disease detection is proposed: local, semi-global, global, and symptom-adaptive annotation.

Hyperspectral machine-learning model for screening tea germplasm resources with drought tolerance. →

Drought tolerance and quality stability are important indicators to evaluate the stress tolerance of tea germplasm resources. The traditional screening method of drought resistant germplasm is mainly to evaluate by detecting physiological and biochemical indicators of tea plants under drought stresses. However, the methods are not only time consuming but also destructive. In this study, hyperspectral images of tea drought phenotypes were obtained and modeled with related physiological indicators. The results showed that: (1) the information contents of malondialdehyde, soluble sugar and total polyphenol were 0.21, 0.209 and 0.227 respectively, and the drought tolerance coefficient (DTC) index of each tea variety was between 0.069 and 0.81; (2) the comprehensive drought tolerance of different varieties were (from strong to weak): QN36, SCZ, ZC108, JX, JGY, XY10, QN1, MS9, QN38 , and QN21 ; (3) by using SVM, RF and PLSR to model DTC (drought tolerance coefficient) data, the best prediction model was selected as MSC-2D-UVE-SVM (R 2 = 0.77, RMSE = 0.073, MAPE = 0.16) for drought tolerance of tea germplasm resources, named Tea-DTC model. Therefore, the Tea-DTC model based on hyperspectral machine-learning technology can be used as a new screening method for evaluating tea germplasm resources with drought tolerance.

Why it matches plant phenotyping methods茶樹の乾燥耐性という植物状態をハイパースペクトル画像と機械学習で推定するモデルを開発し、従来の生理・生化学指標に代わるスクリーニング手法として性能評価しているため、フェノタイピング手法が中心である。

abstractIn this study, hyperspectral images of tea drought phenotypes were obtained and modeled with related physiological indicators.

Characterization and genetic dissection of maize ear leaf midrib acquired by 3D digital technology. →

The spatial morphological structure of plant leaves is an important index to evaluate crop ideotype. In this study, we characterized the three-dimensional (3D) data of the ear leaf midrib of maize at the grain-filling stage using the 3D digitization technology and obtained the phenotypic values of 15 traits covering four different dimensions of the ear leaf midrib, of which 13 phenotypic traits were firstly proposed for featuring plant leaf spatial structure. Cluster analysis results showed that the 13 traits could be divided into four groups, Group I, -II, -III and -IV. Group I contains HorizontalLength, OutwardGrowthMeasure, LeafAngle and DeviationTip; Group II contains DeviationAngle, MaxCurvature and CurvaturePos; Group III contains LeafLength and ProjectionArea; Group IV contains TipTop, VerticalHeight, UpwardGrowthMeasure, and CurvatureRatio. To investigate the genetic basis of the ear leaf midrib curve, 13 traits with high repeatability were subjected to genome-wide association study (GWAS) analysis. A total of 828 significantly related SNPs were identified and 1365 candidate genes were annotated. Among these, 29 candidate genes with the highest significant and multi-method validation were regarded as the key findings. In addition, pathway enrichment analysis was performed on the candidate genes of traits to explore the potential genetic mechanism of leaf midrib curve phenotype formation. These results not only contribute to further understanding of maize leaf spatial structure traits but also provide new genetic loci for maize leaf spatial structure to improve the plant type of maize varieties.

Why it matches plant phenotyping methodsトウモロコシ葉脈の3Dデジタル化により空間形態の表現型形質を取得・定量化することが研究の中心であり、GWASだけでなく新規形質の抽出を扱っている。

abstractwe characterized the three-dimensional (3D) data of the ear leaf midrib of maize at the grain-filling stage using the 3D digitization technology and obtained the phenotypic values of 15 traits covering four different dimensions of the ear leaf midrib

Transfer learning for versatile plant disease recognition with limited data. →

Deep learning has witnessed a significant improvement in recent years to recognize plant diseases by observing their corresponding images. To have a decent performance, current deep learning models tend to require a large-scale dataset. However, collecting a dataset is expensive and time-consuming. Hence, the limited data is one of the main challenges to getting the desired recognition accuracy. Although transfer learning is heavily discussed and verified as an effective and efficient method to mitigate the challenge, most proposed methods focus on one or two specific datasets. In this paper, we propose a novel transfer learning strategy to have a high performance for versatile plant disease recognition , on multiple plant disease datasets. Our transfer learning strategy differs from the current popular one due to the following factors. First, PlantCLEF2022, a large-scale dataset related to plants with 2,885,052 images and 80,000 classes, is utilized to pre-train a model. Second, we adopt a vision transformer (ViT) model, instead of a convolution neural network. Third, the ViT model undergoes transfer learning twice to save computations. Fourth, the model is first pre-trained in ImageNet with a self-supervised loss function and with a supervised loss function in PlantCLEF2022. We apply our method to 12 plant disease datasets and the experimental results suggest that our method surpasses the popular one by a clear margin for different dataset settings. Specifically, our proposed method achieves a mean testing accuracy of 86.29over the 12 datasets in a 20-shot case, 12.76 higher than the current state-of-the-art method's accuracy of 73.53. Furthermore, our method outperforms other methods in one plant growth stage prediction and the one weed recognition dataset. To encourage the community and related applications, we have made public our codes and pre-trained model.

Why it matches plant phenotyping methods植物病害画像から病徴・病害状態を認識するための転移学習手法を開発し、12の病害データセットで性能比較・検証しており、植物フェノタイピング手法が中心である。

abstractwe propose a novel transfer learning strategy to have a high performance for versatile plant disease recognition , on multiple plant disease datasets.

A graph-based approach for simultaneous semantic and instance segmentation of plant 3D point clouds →

Accurate simultaneous semantic and instance segmentation of a plant 3D point cloud is critical for automatic plant phenotyping. Classically, each organ of the plant is detected based on the local geometry of the point cloud, but the consistency of the global structure of the plant is rarely assessed. We propose a two-level, graph-based approach for the automatic, fast and accurate segmentation of a plant into each of its organs with structural guarantees. We compute local geometric and spectral features on a neighbourhood graph of the points to distinguish between linear organs (main stem, branches, petioles) and two-dimensional ones (leaf blades) and even 3-dimensional ones (apices). Then a quotient graph connecting each detected macroscopic organ to its neighbors is used both to refine the labelling of the organs and to check the overall consistency of the segmentation. A refinement loop allows to correct segmentation defects. The method is assessed on both synthetic and real 3D point-cloud data sets of Chenopodium album (wild spinach) and Solanum lycopersicum (tomato plant).

Why it matches plant phenotyping methods植物3D点群から器官を自動分割・識別するグラフベース手法を開発し、合成および実データで評価しており、植物表現型取得の技術が中心である。

abstractAccurate simultaneous semantic and instance segmentation of a plant 3D point cloud is critical for automatic plant phenotyping.

Rapid nondestructive detection of peanut varieties and peanut mildew based on hyperspectral imaging and stacked machine learning models. →

Moldy peanut seeds are damaged by mold, which seriously affects the germination rate of peanut seeds. At the same time, the quality and variety purity of peanut seeds profoundly affect the final yield of peanuts and the economic benefits of farmers. In this study, hyperspectral imaging technology was used to achieve variety classification and mold detection of peanut seeds. In addition, this paper proposed to use median filtering (MF) to preprocess hyperspectral data, use four variable selection methods to obtain characteristic wavelengths, and ensemble learning models (SEL) as a stable classification model. This paper compared the model performance of SEL and extreme gradient boosting algorithm (XGBoost), light gradient boosting algorithm (LightGBM), and type boosting algorithm (CatBoost). The results showed that the MF-LightGBM-SEL model based on hyperspectral data achieves the best performance. Its prediction accuracy on the data training and data testing reach 98.63% and 98.03%, respectively, and the modeling time was only 0.37s, which proved that the potential of the model to be used in practice. The approach of SEL combined with hyperspectral imaging techniques facilitates the development of a real-time detection system. It could perform fast and non-destructive high-precision classification of peanut seed varieties and moldy peanuts, which was of great significance for improving crop yields.

Why it matches plant phenotyping methodsピーナッツ種子の品種分類とカビ状態検出を目的に、ハイパースペクトル画像と前処理・機械学習モデルを中心的に開発・比較しており、植物の状態を推定するフェノタイピング手法に該当する。

abstracthyperspectral imaging technology was used to achieve variety classification and mold detection of peanut seeds.

A comparison of classical and machine learning-based phenotype prediction methods on simulated data and three plant species →

Genomic selection is an integral tool for breeders to accurately select plants directly from genotype data leading to faster and more resource-efficient breeding programs. Several prediction methods have been established in the last few years. These range from classical linear mixed models to complex non-linear machine learning approaches, such as Support Vector Regression, and modern deep learning-based architectures. Many of these methods have been extensively evaluated on different crop species with varying outcomes. In this work, our aim is to systematically compare 12 different phenotype prediction models, including basic genomic selection methods to more advanced deep learning-based techniques. More importantly, we assess the performance of these models on simulated phenotype data as well as on real-world data from Arabidopsis thaliana and two breeding datasets from soy and corn. The synthetic phenotypic data allow us to analyze all prediction models and especially the selected markers under controlled and predefined settings. We show that Bayes B and linear regression models with sparsity constraints perform best under different simulation settings with respect to explained variance. Further, we can confirm results from other studies that there is no superiority of more complex neural network-based architectures for phenotype prediction compared to well-established methods. However, on real-world data, for which several prediction models yield comparable results with slight advantages for Elastic Net, this picture is less clear, suggesting that there is a lot of room for future research.

Why it matches plant phenotyping methods複数の表現型予測モデルを植物種の実データとシミュレーションで系統比較・評価しており、計算による植物形質推定が研究の中心である。

abstractour aim is to systematically compare 12 different phenotype prediction models

Climate change conditions the selection of rust-resistant candidate wild lentil populations for in situ conservation. →

Crop Wild Relatives (CWR) are a valuable source of genetic diversity that can be transferred to commercial crops, so their conservation will become a priority in the face of climate change. Bizarrely, in situ conserved CWR populations and the traits one might wish to preserve in them are themselves vulnerable to climate change. In this study, we used a quantitative machine learning predictive approach to project the resistance of CWR populations of lentils to a common disease, lentil rust, caused by fungus Uromyces viciae-fabae . Resistance is measured through a proxy quantitative value, DSr (Disease Severity relative), quite complex and expensive to get. Therefore, machine learning is a convenient tool to predict this magnitude using a well-curated georeferenced calibration set. Previous works have provided a binary outcome (resistant vs. non-resistant), but that approach is not fine enough to answer three practical questions: which variables are key to predict rust resistance, which CWR populations are resistant to rust under current environmental conditions, and which of them are likely to keep this trait under different climate change scenarios. We first predict rust resistance in present time for crop wild relatives that grow up inside protected areas. Then, we use the same models under future climate IPCC (Intergovernmental Panel on Climate Change) scenarios to predict future DSr values. Populations that are rust-resistant by now and under future conditions are optimal candidates for further evaluation and in situ conservation of this valuable trait. We have found that rust-resistance variation as a result of climate change is not uniform across the geographic scope of the study (the Mediterranean basin), and that candidate populations share some interesting common environmental conditions.

Why it matches plant phenotyping methods機械学習により、取得が困難なレンチルの病害重症度（DSr）という植物の病害状態を定量予測する手法が研究の中心であり、現況・将来条件での適用も評価している。

abstractwe used a quantitative machine learning predictive approach to project the resistance of CWR populations of lentils to a common disease, lentil rust

A weight optimization-based transfer learning approach for plant disease detection of New Zealand vegetables. →

Deep learning (DL) is an effective approach to identifying plant diseases. Among several DL-based techniques, transfer learning (TL) produces significant results in terms of improved accuracy. However, the usefulness of TL has not yet been explored using weights optimized from agricultural datasets. Furthermore, the detection of plant diseases in different organs of various vegetables has not yet been performed using a trained/optimized DL model. Moreover, the presence/detection of multiple diseases in vegetable organs has not yet been investigated. To address these research gaps, a new dataset named NZDLPlantDisease-v2 has been collected for New Zealand vegetables. The dataset includes 28 healthy and defective organs of beans, broccoli, cabbage, cauliflower, kumara, peas, potato, and tomato. This paper presents a transfer learning method that optimizes weights obtained through agricultural datasets for better outcomes in plant disease identification. First, several DL architectures are compared to obtain the best-suited model, and then, data augmentation techniques are applied. The Faster Region-based Convolutional Neural Network (RCNN) Inception ResNet-v2 attained the highest mean average precision (mAP) compared to the other DL models including different versions of Faster RCNN, Single-Shot Multibox Detector (SSD), Region-based Fully Convolutional Networks (RFCN), RetinaNet, and EfficientDet. Next, weight optimization is performed on datasets including PlantVillage, NZDLPlantDisease-v1, and DeepWeeds using image resizers, interpolators, initializers, batch normalization, and DL optimizers. Updated/optimized weights are then used to retrain the Faster RCNN Inception ResNet-v2 model on the proposed dataset. Finally, the results are compared with the model trained/optimized using a large dataset, such as Common Objects in Context (COCO). The final mAP improves by 9.25% and is found to be 91.33%. Moreover, the robustness of the methodology is demonstrated by testing the final model on an external dataset and using the stratified k-fold cross-validation method.

Why it matches plant phenotyping methods植物病害の症状を画像から検出する深層学習手法を開発し、データセット、モデル比較、外部検証、交差検証まで行っており、植物状態の取得・推定が中心である。

abstractThis paper presents a transfer learning method that optimizes weights obtained through agricultural datasets for better outcomes in plant disease identification.

Rapid prediction of winter wheat yield and nitrogen use efficiency using consumer-grade unmanned aerial vehicles multispectral imagery. →

Rapid and accurate assessment of yield and nitrogen use efficiency (NUE) is essential for growth monitoring, efficient utilization of fertilizer and precision management. This study explored the potential of a consumer-grade DJI Phantom 4 Multispectral (P4M) camera for yield or NUE assessment in winter wheat by using the universal vegetation indices independent of growth period. Three vegetation indices having a strong correlation with yield or NUE during the entire growth season were determined through Pearson's correlational analysis, while multiple linear regression (MLR), stepwise MLR (SMLR), and partial least-squares regression (PLSR) methods based on the aforementioned vegetation indices were adopted during different growth periods. The cumulative results showed that the reciprocal ratio vegetation index (repRVI) had a high potential for yield assessment throughout the growing season, and the late grain-filling stage was deemed as the optimal single stage with R 2 , root mean square error (RMSE), and mean absolute error (MAE) of 0.85, 793.96 kg/ha, and 656.31 kg/ha, respectively. MERIS terrestrial chlorophyll index (MTCI) performed better in the vegetative period and provided the best prediction results for the N partial factor productivity (NPFP) at the jointing stage, with R 2 , RMSE, and MAE of 0.65, 10.53 kg yield/kg N, and 8.90 kg yield/kg N, respectively. At the same time, the modified normalized difference blue index (mNDblue) was more accurate during the reproductive period, providing the best accuracy for agronomical NUE (aNUE) assessment at the late grain-filling stage, with R 2 , RMSE, and MAE of 0.61, 7.48 kg yield/kg N, and 6.05 kg yield/kg N, respectively. Furthermore, the findings indicated that model accuracy cannot be improved by increasing the number of input features. Overall, these results indicate that the consumer-grade P4M camera is suitable for early and efficient monitoring of important crop traits, providing a cost-effective choice for the development of the precision agricultural system.

Why it matches plant phenotyping methods消費者向けUAVマルチスペクトル画像を用いて小麦の収量および窒素利用効率を推定・検証する方法が研究の中心であり、植物形質の取得と予測性能を評価している。

abstractThis study explored the potential of a consumer-grade DJI Phantom 4 Multispectral (P4M) camera for yield or NUE assessment in winter wheat

Efficient attention-based CNN network (EANet) for multi-class maize crop disease classification →

Maize leaf disease significantly reduces the quality and overall crop yield. Therefore, it is crucial to monitor and diagnose illnesses during the growth season to take necessary actions. However, accurate identification is challenging to achieve as the existing automated methods are computationally complex or perform well on images with a simple background. Whereas, the realistic field conditions include a lot of background noise that makes this task difficult. In this study, we presented an end-to-end learning CNN architecture, Efficient Attention Network (EANet) based on the EfficientNetv2 model to identify multi-class maize crop diseases. To further enhance the capacity of the feature representation, we introduced a spatial-channel attention mechanism to focus on affected locations and help the detection network accurately recognize multiple diseases. We trained the EANet model using focal loss to overcome class-imbalanced data issues and transfer learning to enhance network generalization. We evaluated the presented approach on the publically available datasets having samples captured under various challenging environmental conditions such as varying background, non-uniform light, and chrominance variances. Our approach showed an overall accuracy of 99.89% for the categorization of various maize crop diseases. The experimental and visual findings reveal that our model shows improved performance compared to conventional CNNs, and the attention mechanism properly accentuates the disease-relevant information by ignoring the background noise.

Why it matches plant phenotyping methodsトウモロコシ葉の画像から病害状態を分類するCNN手法を開発・評価しており、植物病害表現型の取得・抽出が研究の中心である。

abstractwe presented an end-to-end learning CNN architecture, Efficient Attention Network (EANet) based on the EfficientNetv2 model to identify multi-class maize crop diseases.

Characterization of complex photosynthetic pigment profiles in European deciduous tree leaves by sequential extraction and reversed-phase high-performance liquid chromatography. →

Leaf pigments, including chlorophylls and carotenoids, are important biochemical indicators of plant photosynthesis and photoprotection. In this study, we developed, optimized, and validated a sequential extraction and liquid chromatography-diode array detection method allowing for the simultaneous quantification of the main photosynthetic pigments, including chlorophyll a, chlorophyll b, β-carotene, lutein, neoxanthin, and the xanthophyll cycle (VAZ), as well as the characterization of plant pigment derivatives. Chromatographic separation was accomplished with the newest generation of core-shell columns revealing numerous pigment derivatives. The sequential extraction allowed for a better recovery of the main pigments (+25 % chlorophyll a, +30 % chlorophyll b, +42 % β-carotene, and 61% xanthophylls), and the characterization of ca. 5.3 times more pigment derivatives (i.e., up to 62 chlorophyll and carotenoid derivatives including isomers) than with a single-step extraction. A broad working range of concentrations (300-2,000 ng.mL -1 ) was achieved for most pigments and their derivatives and the limit of detection was as low as a few nanograms per milliliter. The method also showed adequate trueness (RSD Fagus sylvatica L . leaves, pigment derivatives revealed a high within-individual tree variability throughout the growing season that could not be detected using the main photosynthetic pigments alone, eventually showing that the method allowed for the monitoring of pigment dynamics at unprecedented detail.

Why it matches plant phenotyping methods葉の光合成色素を植物の生理状態・動態として定量する分析法を開発、最適化、検証しており、表現型取得法が研究の中心である。

abstractIn this study, we developed, optimized, and validated a sequential extraction and liquid chromatography-diode array detection method allowing for the simultaneous quantification of the main photosynthetic pigments

Deep learning-based segmentation and classification of leaf images for detection of tomato plant disease →

Plants contribute significantly to the global food supply. Various Plant diseases can result in production losses, which can be avoided by maintaining vigilance. However, manually monitoring plant diseases by agriculture experts and botanists is time-consuming, challenging and error-prone. To reduce the risk of disease severity, machine vision technology (i.e., artificial intelligence) can play a significant role. In the alternative method, the severity of the disease can be diminished through computer technologies and the cooperation of humans. These methods can also eliminate the disadvantages of manual observation. In this work, we proposed a solution to detect tomato plant disease using a deep leaning-based system utilizing the plant leaves image data. We utilized an architecture for deep learning based on a recently developed convolutional neural network that is trained over 18,161 segmented and non-segmented tomato leaf images-using a supervised learning approach to detect and recognize various tomato diseases using the Inception Net model in the research work. For the detection and segmentation of disease-affected regions, two state-of-the-art semantic segmentation models, i.e., U-Net and Modified U-Net, are utilized in this work. The plant leaf pixels are binary and classified by the model as Region of Interest (ROI) and background. There is also an examination of the presentation of binary arrangement (healthy and diseased leaves), six-level classification (healthy and other ailing leaf groups), and ten-level classification (healthy and other types of ailing leaves) models. The Modified U-net segmentation model outperforms the simple U-net segmentation model by 98.66 percent, 98.5 IoU score, and 98.73 percent on the dice. InceptionNet1 achieves 99.95% accuracy for binary classification problems and 99.12% for classifying six segmented class images; InceptionNet outperformed the Modified U-net model to achieve higher accuracy. The experimental results of our proposed method for classifying plant diseases demonstrate that it outperforms the methods currently available in the literature.

Why it matches plant phenotyping methodsトマト葉画像から病斑領域をセグメンテーションし、植物病害を分類する深層学習手法の開発・評価が研究の中心であるため、植物表現型計測手法として含める。

abstractIn this work, we proposed a solution to detect tomato plant disease using a deep leaning-based system utilizing the plant leaves image data.

JustDeepIt: Software tool with graphical and character user interfaces for deep learning-based object detection and segmentation in image analysis. →

Image processing and analysis based on deep learning are becoming mainstream and increasingly accessible for solving various scientific problems in diverse fields. However, it requires advanced computer programming skills and a basic familiarity with character user interfaces (CUIs). Consequently, programming beginners face a considerable technical hurdle. Because potential users of image analysis are experimentalists, who often use graphical user interfaces (GUIs) in their daily work, there is a need to develop GUI-based easy-to-use deep learning software to support their work. Here, we introduce JustDeepIt, a software written in Python, to simplify object detection and instance segmentation using deep learning. JustDeepIt provides both a GUI and a CUI. It contains various functional modules for model building and inference, and it is built upon the popular PyTorch, MMDetection, and Detectron2 libraries. The GUI is implemented using the Python library FastAPI, simplifying model building for various deep learning approaches for beginners. As practical examples of JustDeepIt, we prepared four case studies that cover critical issues in plant science: (1) wheat head detection with Faster R-CNN, YOLOv3, SSD, and RetinaNet; (2) sugar beet and weed segmentation with Mask R-CNN; (3) plant segmentation with U 2 -Net; and (4) leaf segmentation with U 2 -Net. The results support the wide applicability of JustDeepIt in plant science applications. In addition, we believe that JustDeepIt has the potential to be applied to deep learning-based image analysis in various fields beyond plant science.

Why it matches plant phenotyping methods植物画像の物体検出・インスタンスセグメンテーションを行うソフトウェア自体が中心で、植物科学での検証例も含むため、植物フェノタイピング手法として含める。

abstractHere, we introduce JustDeepIt, a software written in Python, to simplify object detection and instance segmentation using deep learning.

DIANA: A deep learning-based paprika plant disease and pest phenotyping system with disease severity analysis →

The emergence of deep neural networks has allowed the development of fully automated and efficient diagnostic systems for plant disease and pest phenotyping. Although previous approaches have proven to be promising, they are limited, especially in real-life scenarios, to properly diagnose and characterize the problem. In this work, we propose a framework which besides recognizing and localizing various plant abnormalities also informs the user about the severity of the diseases infecting the plant. By taking a single image as input, our algorithm is able to generate detailed descriptive phrases (user-defined) that display the location, severity stage, and visual attributes of all the abnormalities that are present in the image. Our framework is composed of three main components. One of them is a detector that accurately and efficiently recognizes and localizes the abnormalities in plants by extracting region-based anomaly features using a deep neural network-based feature extractor. The second one is an encoder-decoder network that performs pixel-level analysis to generate abnormality-specific severity levels. Lastly is an integration unit which aggregates the information of these units and assigns unique IDs to all the detected anomaly instances, thus generating descriptive sentences describing the location, severity, and class of anomalies infecting plants. We discuss two possible ways of utilizing the abovementioned units in a single framework. We evaluate and analyze the efficacy of both approaches on newly constructed diverse paprika disease and pest recognition datasets, comprising six anomaly categories along with 11 different severity levels. Our algorithm achieves mean average precision of 91.7% for the abnormality detection task and a mean panoptic quality score of 70.78% for severity level prediction. Our algorithm provides a practical and cost-efficient solution to farmers that facilitates proper handling of crops.

Why it matches plant phenotyping methods画像から植物病害・害虫の位置と重症度を推定する深層学習システムの開発・評価が研究の中心であり、植物の病害状態を直接定量化するため、植物フェノタイピング手法として含める。

abstractwe propose a framework which besides recognizing and localizing various plant abnormalities also informs the user about the severity of the diseases infecting the plant.

4DRoot: Root phenotyping software for temporal 3D scans by X-ray computed tomography →

Currently, plant phenomics is considered the key to reducing the genotype-to-phenotype knowledge gap in plant breeding. In this context, breakthrough imaging technologies have demonstrated high accuracy and reliability. The X-ray computed tomography (CT) technology can noninvasively scan roots in 3D; however, it is urgently required to implement high-throughput phenotyping procedures and analyses to increase the amount of data to measure more complex root phenotypic traits. We have developed a spatial-temporal root architectural modeling software tool based on 4D data from temporal X-ray CT scans. Through a cylinder fitting, we automatically extract significant root architectural traits, distribution, and hierarchy. The open-source software tool is named 4DRoot and implemented in MATLAB. The source code is freely available at https://github.com/TIDOP-USAL/4DRoot. In this research, 3D root scans from the black walnut tree were analyzed, a punctual scan for the spatial study and a weekly time-slot series for the temporal one. 4DRoot provides breeders and root biologists an objective and useful tool to quantify carbon sequestration throw trait extraction. In addition, 4DRoot could help plant breeders to improve plants to meet the food, fuel, and fiber demands in the future, in order to increase crop yield while reducing farming inputs.

Why it matches plant phenotyping methodsX線CTの時系列3D画像から根系形態形質を自動抽出するソフトウェア開発が研究の中心であり、植物フェノタイピング手法に該当する。

abstractWe have developed a spatial-temporal root architectural modeling software tool based on 4D data from temporal X-ray CT scans.

Multi-environment genome -wide association mapping of culm morphology traits in barley. →

In cereals with hollow internodes, lodging resistance is influenced by morphological characteristics such as internode diameter and culm wall thickness. Despite their relevance, knowledge of the genetic control of these traits and their relationship with lodging is lacking in temperate cereals such as barley. To fill this gap, we developed an image analysis-based protocol to accurately phenotype culm diameters and culm wall thickness across 261 barley accessions. Analysis of culm trait data collected from field trials in seven different environments revealed high heritability values (>50%) for most traits except thickness and stiffness, as well as genotype-by-environment interactions. The collection was structured mainly according to row-type, which had a confounding effect on culm traits as evidenced by phenotypic correlations. Within both row-type subsets, outer diameter and section modulus showed significant negative correlations with lodging (<-0.52 and <-0.45, respectively), but no correlation with plant height, indicating the possibility of improving lodging resistance independent of plant height. Using 50k iSelect SNP genotyping data, we conducted multi-environment genome-wide association studies using mixed model approach across the whole panel and row-type subsets: we identified a total of 192 quantitative trait loci (QTLs) for the studied traits, including subpopulation-specific QTLs and 21 main effect loci for culm diameter and/or section modulus showing effects on lodging without impacting plant height. Providing insights into the genetic architecture of culm morphology in barley and the possible role of candidate genes involved in hormone and cell wall-related pathways, this work supports the potential of loci underpinning culm features to improve lodging resistance and increase barley yield stability under changing environments.

Why it matches plant phenotyping methods大麦の穂軸径と穂軸壁厚を測定する画像解析プロトコルを開発し、多数アクセッション・複数環境で適用しており、表現型取得法が研究の中心的要素である。

abstractwe developed an image analysis-based protocol to accurately phenotype culm diameters and culm wall thickness across 261 barley accessions.

A workflow for segmenting soil and plant X-ray computed tomography images with deep learning in Google’s Colaboratory →

X-ray micro-computed tomography (X-ray μCT) has enabled the characterization of the properties and processes that take place in plants and soils at the micron scale. Despite the widespread use of this advanced technique, major limitations in both hardware and software limit the speed and accuracy of image processing and data analysis. Recent advances in machine learning, specifically the application of convolutional neural networks to image analysis, have enabled rapid and accurate segmentation of image data. Yet, challenges remain in applying convolutional neural networks to the analysis of environmentally and agriculturally relevant images. Specifically, there is a disconnect between the computer scientists and engineers, who build these AI/ML tools, and the potential end users in agricultural research, who may be unsure of how to apply these tools in their work. Additionally, the computing resources required for training and applying deep learning models are unique, more common to computer gaming systems or graphics design work, than to traditional computational systems. To navigate these challenges, we developed a modular workflow for applying convolutional neural networks to X-ray μCT images, using low-cost resources in Google's Colaboratory web application. Here we present the results of the workflow, illustrating how parameters can be optimized to achieve best results using example scans from walnut leaves, almond flower buds, and a soil aggregate. We expect that this framework will accelerate the adoption and use of emerging deep learning techniques within the plant and soil sciences.

Why it matches plant phenotyping methods植物試料のX線μCT画像を対象に、CNNによるセグメンテーション workflow を開発・最適化しており、植物画像からの表現型情報抽出法が中心である。

abstractwe developed a modular workflow for applying convolutional neural networks to X-ray μCT images, using low-cost resources in Google's Colaboratory web application.

Flexible and high quality plant growth prediction with limited data. →