Pearl millet is an important crop in arid regions, but its yield is reduced by foliar diseases like Downy Mildew and Rust. Traditional and deep learning methods struggle with accurate lesion detection, severity estimation, and robustness under complex field conditions, and often lack interpretability for practical agricultural deployment. To address these challenges, this study proposes the Adaptive Severity-Aware Swin Attention Network (ASA-SAN), an integrated framework designed for disease segmentation, classification, and severity estimation in pearl millet leaves. The proposed architecture combines a Swin Transformer encoder for hierarchical feature extraction with a ResUNet++ decoder for accurate lesion segmentation. This is further enhanced using Adaptive Channel Attention to improve feature discrimination and a dual-stream classification network to jointly capture local lesion characteristics and global contextual information. Additionally, an Adaptive Disease Severity Index (ADSI) is introduced to quantitatively assess disease progression based on lesion area ratio, color degradation, edge irregularity, and texture variations. Experimental evaluations conducted on a pearl millet leaf dataset demonstrate that the proposed method achieves a Dice score of 97.8%, IoU of 95.6%, classification accuracy of 98.3%, and F1-score of 98.2%, outperforming several state-of-the-art methods. Furthermore, Grad-CAM visualizations enhance model interpretability by highlighting disease-relevant regions. Overall, the ASA-SAN framework provides a robust, interpretable, and severity-aware solution for automated pearl millet disease analysis, enabling early detection and supporting precision agriculture practices for improved crop protection and yield optimization.
Why it matches plant phenotyping methods真珠粟葉の病斑を画像から分割・分類し、病害重症度を定量推定する手法を中心に開発・評価しているため、植物表現型計測手法として含める。
abstractAdditionally, an Adaptive Disease Severity Index (ADSI) is introduced to quantitatively assess disease progression based on lesion area ratio, color degradation, edge irregularity, and texture variations.
Reproduction assets foundThe paper's phenotyping inputs are drawn from a public, open-access image dataset: the Pearl Millet Leaf Disease dataset (Version 2) hosted on Roboflow Universe, containing annotated images of Downy Mildew, Rust, and healthy pearl millet leaves. This is a paper-specific, publicly available asset directly used for the作者Dataset · publicThe dataset used in this research was taken from the publicly available open-access Pearl
Millet Leaf Disease dataset hosted on Roboflow Universe, which has images of Downy
Mildew, Rust and healthy pearl millet leaves annotated publicly available [26]. To ensure
experimental consistency and reproducibility, all experiments were conducted with Version
2 of the open access dataset.Open asset ↗Roboflow Universepdf-raw-page:10 lines:1-28Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
Pearl millet is a vital crop in arid and semi-arid regions, but its productivity is significantly impacted by fungal and bacterial diseases. Traditional disease detection methods, including manual inspection and conventional machine learning techniques, often suffer from inefficiencies due to subjectivity, time consumption, and limited feature extraction capabilities. Deep learning-based segmentation models such as U-Net, SegNet, DeepLabV3, and FCN have been employed to automate disease identification, but they exhibit limitations in accurately localizing disease-affected regions, handling complex spatial patterns, and maintaining high segmentation precision. To address these challenges, this study proposes HAMNet-Mask R-CNN, an AI-powered hierarchical multi-scale attention network integrated with a spectral-spatial fusion module for precise pearl millet disease detection. The novel framework enhances disease localization by utilizing hierarchical multi-scale feature learning, where low, mid, and high-level features contribute to superior classification accuracy. The attention mechanism prioritizes critical disease regions, reducing false positives, while Mask R-CNN enables pixel-wise segmentation, refining disease boundary identification. The proposed model is implemented using Python and TensorFlow with high-resolution image datasets. Performance evaluation demonstrates HAMNet-Mask R-CNN achieving a 99.35% Dice score, 98.80% IoU, 99.75% Precision, 99.78% Recall, and 99.65% Accuracy, outperforming U-Net (95.03% Dice, 90.53% IoU), SegNet (94.58% Dice, 89.73% IoU), DeepLabV3 (98.31% Dice, 96.69% IoU) and FCN(98.47% Dice score, 97.00% IoU). The results confirm the model's superiority in disease segmentation and classification, offering a robust and scalable solution for real-time disease monitoring in smart agriculture. The integration of hierarchical attention and spectral-spatial fusion significantly enhances detection accuracy, ensuring reliable disease identification and contributing to improved agricultural productivity.
Why it matches plant phenotyping methods真珠粟の病変領域を画像から画素単位で抽出・分類する深層学習手法を開発し、既存モデルと性能比較しているため、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 proposes HAMNet-Mask R-CNN, an AI-powered hierarchical multi-scale attention network integrated with a spectral-spatial fusion module for precise pearl millet disease detection.
Reproduction assets foundThe paper's authors explicitly state that the custom HAMNet–Mask R-CNN implementation code (preprocessing, training, evaluation, visualization) is publicly available on GitHub. The pearl millet leaf image dataset used as phenotyping input is also publicly hosted on Roboflow Universe and cited as the data source. The roCode · publicThe custom code developed for the implementation of the proposed HAMNet–Mask R-CNN framework, including data preprocessing, model training, evaluation, and visualization scripts, is publicly available in a GitHub repository. The code can be accessed at: https://github.com/ramyalaksha/Hamnet.gitOpen asset ↗https://github.com/ramyalaksha/Hamnet.gitlines:267-297Code / dataset availability confirmedEurope PMC · checked 15 Sept 2026
Published28 Mar 2026Metabolomics : Official journal of the Metabolomic SocietyCited by 0 · OpenAlex ↗
Introduction Pearl millet is a high nutritional cereal recognised for its agro-climatic resilience, making it relevant for food security under climate change scenarios. Phenotypic traits are indicative of crop performance, stability and adaptability, yet the potential of metabolomics to predict these traits has not been explored. Objectives This study aimed to identify metabolite-trait associations in the Brazilian germplasm core collection, comprising 203 pearl millet genotypes, by combining untargeted metabolomics with machine-learning models. Methods Grains metabolic profiles were obtained using untargeted UHPLC-LTQ-Orbitrap-HRMS. Phenotypic data were sourced from standardised evaluations conducted by Embrapa across different years and field trials within the Sete Lagoas experimental station (Minas Gerais, Brazil). Generalised linear modelling with penalisation (GLM) and Random Forest was applied to explore the correlation between metabolism and 21 phenotypic traits. Results GLM successfully predicted eight qualitative and seven quantitative traits. Prediction accuracy was higher for qualitative traits, reflecting their comparatively simpler genetic architecture, whereas quantitative traits also achieved satisfactory performance (R² ≥ 0.6). Key predictors included phenolic compounds, amino acids, fatty acids, and carbohydrates. Notably, several associations corresponded to metabolites involved in nitrogen metabolism and vegetative growth, underscoring biologically meaningful links between metabolic profiles and trait variation. Conclusions This exploratory study presents the first metabolome characterisation of a pearl millet germplasm bank, coupled with predictive modelling of phenotypic traits. However, our findings are constrained by the single-environment design and the absence of population-structure assessment. To establish the stability and biological relevance of these results, future work should incorporate multi-environment trials and pathway-level analyses accounting for population structure.
Why it matches plant phenotyping methodsメタボロームを入力として機械学習で植物の表現型形質を予測し、複数形質で予測精度を評価しているため、単なる生物学的測定ではなく形質推定手法の検証が中心です。
abstractThis study aimed to identify metabolite-trait associations in the Brazilian germplasm core collection, comprising 203 pearl millet genotypes, by combining untargeted metabolomics with machine-learning models.
Reproduction assets foundThe paper deposits its metabolomics and phenotypic metadata in a public repository (Recherche Data Gouv, DOI 10.57745/GU6WDG). No author analysis code or trained model deposit is stated; supplementary materials are not linked to a qualifying URL.Dataset · public.623/2023; 26/210.152/2023; 26/201.317/2022), National Council for Scientific and Technological Development (CNPq) (407350/2023-3; 314100/2023-7), Coordination for Improvement of Personnel with Higher Education (CAPES) (financial code 001).
Data availability
The metabolomics and metadata reported in this paper are available via https://doi.org/10.57745/GU6WDG.
Declarations
Competing interests
The authors declare no competing interests.
References
Alonso-Blanco C Méndez-Vigo B
Genetic architecture of naturally occurring quantitative traits in plants: An updated synthesis
Current Opinion in Plant Biology 2014 18 37 43
10.1016/j.pbi.2014.01.002
24565952
Alonso-Blanco, C., & Méndez-VigoOpen asset ↗10.57745 · GU6WDGlines:121-160Code / dataset availability confirmedEurope PMC · checked 5 Sept 2026
MilletSorghumRootTissueSegmentationRoot system architecture
Root anatomical features are critical for plant performance characterization, yet phenotyping at the anatomical scale remains limited by the extreme annotation burden of cellular segmentation. We present a two-stage segmentation framework that greatly reduces annotation requirements while maintaining high accuracy across diverse plant species and imaging conditions. Our approach decomposes multi-class segmentation into species-agnostic tissue identification followed by tissue type classification. By designing robust input representations invariant to imaging artifacts and morphological variations, our framework enables rapid adaptation to new species with fewer than 40 labeled images. Additionally, the first stage automatically generates tissue boundaries, transforming tedious manual tracing into simple tissue labeling. We validate our method on pearl millet, and sorghum root cross-sections from different imaging protocols, achieving state-of-the-art performance while dramatically reducing deployment time. This efficiency breakthrough enables scalable root phenotyping across diverse crop species, accelerating the development of climate-resilient varieties for global food security.
Why it matches plant phenotyping methods植物根の解剖学的形質を対象とする画像セグメンテーション手法を開発し、複数種・撮像条件で検証しているため、方法が研究の中心である。
abstractWe present a two-stage segmentation framework that greatly reduces annotation requirements while maintaining high accuracy across diverse plant species and imaging conditions.
Reproduction assets foundThe paper's Data availability statement explicitly deposits the annotated root image dataset (Zenodo 17726414), trained segmentation models (Zenodo 17737703), and the authors' source code (GitHub janetkok/Root-Segmentation-Beyond-Species-Boundaries), all directly reproducing this paper's root anatomical phenotyping andDataset · publicThe dataset and models are available at https://doi.org/10.5281/zenodo.17726414 and https://doi.org/10.5281/zenodo.17737703 , respectively.Open asset ↗Zenodo · 10.5281/zenodo.17726414lines:242-251Code · publicThe source code is hosted at https://github.com/janetkok/Root-Segmentation-Beyond-Species-Boundaries .Open asset ↗GitHub · janetkok/Root-Segmentation-Beyond-Species-Boundarieslines:242-251Code / dataset availability confirmedEurope PMC · checked 6 Sept 2026
Clustering algorithms are widely used for phenotypic characterization and germplasm management, particularly in data-scarce crops such as neglected and underutilized species (NUS) that lack genomic resources. However, their performance under biologically realistic conditions remains poorly understood. Standard clustering methods commonly applied in crop research often assume distinct, isotropic, and homogeneous clusters, assumptions rarely satisfied in real-world phenotypic datasets. We developed a flexible and empirically calibrated simulation framework, using phenotypic data from West African fonio (Digitaria exilis), to benchmark the performance of eleven clustering algorithms under both idealized and realistic scenarios. Our simulations integrated heterogeneous trait distributions (normal, gamma), strong inter-trait correlations (up to r = -0.84), heteroscedasticity, and moderate population structure (mean Pst = 0.16 ± 0.001, achieved through iterative calibration). Each scenario was replicated 100 times, with clustering accuracy evaluated using external (ARI, NMI) and internal (Silhouette, Davies-Bouldin) validation metrics under standardized conditions. The results revealed consistently poor algorithm performance under realistic conditions (e.g., ARI < 0.07), including for widely used methods in Neglected and Underutilized Species (NUS) research such as K-means, GMM, and PAM. Notably, conventional validation metrics failed to detect biologically meaningful structure revealed by geometric diagnostics, highlighting a critical methodological limitation. Performance markedly improved under idealized conditions, validating our simulation framework. These findings highlight the risk of overinterpreting clustering outputs from weakly structured phenotypic datasets and expose key limitations in current biodiversity analysis practices, particularly those guiding plant genetic resource conservation programs. We provide an open-source R-based diagnostic tool, with parameter specifications to assist practitioners in selecting reproducible and interpretable clustering approaches for germplasm management and biodiversity assessment in data-scarce crops.
Why it matches plant phenotyping methods植物の表現型データを対象に、クラスタリング手法を現実的な条件でベンチマークするシミュレーション枠組みとR診断ツールを開発しており、表現型解析手法が研究の中心である。
abstractWe developed a flexible and empirically calibrated simulation framework, using phenotypic data from West African fonio (Digitaria exilis), to benchmark the performance of eleven clustering algorithms under both idealized and realistic scenarios.
Reproduction assets foundThe paper's Data Availability statement explicitly deposits the complete R simulation/clustering/evaluation script on Zenodo (DOI 10.5281/zenodo.15877863), a paper-specific, publicly actionable code asset. The empirical fonio trait data belong to a prior cited study (Bio et al.), not this paper, and supporting files/DOCode · publicthe complete R script used to simulate phenotypic datasets, apply clustering algorithms, and compute evaluation metrics is publicly available on Zenodo: https://doi.org/10.5281/zenodo.15877863Open asset ↗Zenodo · 10.5281/zenodo.15877863lines:107-122Code / dataset availability confirmedEurope PMC · checked 6 Sept 2026
MaizeMilletRiceSeed / grainMorphology / geometry measurementSegmentationRoot system architecture
Radicle length is a critical indicator of seed vigor, germination capacity, and seedling growth potential. However, existing measurement methods face challenges in automation, efficiency, and generalizability, often requiring manual intervention or re-annotation for different seed types. To address these limitations, this paper proposes an automated method, LenRuler, with a primary focus on rice seeds and validation in multiple crops. The method leverages the Segment Anything Model (SAM) as the foundational segmentation model and employs a coarse-to-fine segmentation strategy combined with Gaussian-based classification to automatically generate bounding boxes and centroids, which are then fed into SAM for precise segmentation of the seed coat and radicle. The radicle length is subsequently computed by converting the geodesic distance between the radicle skeleton's farthest endpoint and its nearest intersection with the seed coat skeleton into the true length. Experiments on the Riceseed1 dataset show that the proposed method achieves a Dice coefficient of 0.955 and a Pixel Accuracy of 0.944, demonstrating excellent segmentation performance. Radicle length measurement experiments on the Riceseed2 test set show that the Mean Absolute Error (MAE) was 0.273 mm and the coefficient of determination (R 2 ) was 0.982, confirming the method's high precision for rice. On the Otherseed dataset, the predicted radicle lengths for maize ( Zea mays ), pearl millet ( Pennisetum glaucum ), and rye ( Secale cereale ) are consistent with the observed radicle length distributions, demonstrating strong cross-species performance. These results establish LenRuler as an accurate and automated solution for radicle length measurement in rice, with validated applicability to other crop species.
Why it matches plant phenotyping methodsイネを中心に複数作物の幼根長を画像から自動抽出・推定する手法を開発し、セグメンテーション性能と測定精度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。
abstractthis paper proposes an automated method, LenRuler, with a primary focus on rice seeds and validation in multiple crops.
Reproduction assets foundThe authors explicitly state that the LenRuler code and software are publicly available on GitHub, providing the paper's radicle-length phenotyping analysis pipeline (SAM-based segmentation, YOLO detection, Gaussian classification).Code · publicThe code and software are available on GitHub at https://github.com/cccccabbage/LenRuler .Open asset ↗cccccabbage/LenRulerlines:351-417Code / dataset availability confirmedOpenAlex · Crossref · Europe PMC · checked 15 Sept 2026
BACKGROUND: graminoid crop cultivated mainly in the arid and semiarid regions of China for more than 7000 years. Its grain highly nutritious and is rich in starch, protein, essential vitamins such as carotenoids, folate, and minerals. To expand the utilisation of foxtail millet, efficient and precise methods for dynamic phenotyping of its growth stages are needed. Traditional foxtail millet monitoring methods have high labour costs and are inefficient and inaccurate, impeding the precise evaluation of foxtail millet genotypic variation. RESULTS: This study introduces a high-throughput imaging system (HIS) with advanced image processing techniques to enhance monitoring efficiency and data quality. The HIS can accurately extract a range of key growth feature parameters, such as plant height (PH), convex hull area (CHA), side projected area (SPA) and colour distribution, from foxtail millet images. Compared with traditional manual measurements, this HIS improved data quality and phenotyping of the key foxtail millet growth traits. High-throughput phenotyping combined with a genome-wide association study (GWAS) revealed genetic loci associated with dynamic growth traits, particularly plant height (PH), in foxtail millet. The loci were linked to genes involved in the gibberellic acid (GA) synthesis pathway related to PH. CONCLUSION: The HIS developed in this study enables the efficient and dynamic monitoring of foxtail millet phenotypic traits. It significantly improves the quality of data obtained for phenotyping key growth traits. The integration of high-throughput phenotyping with GWAS provides new insights into the genetic underpinnings of dynamic growth traits, particularly plant height, by identifying associated genetic loci in the GA synthesis pathway. This methodological advancement opens new avenues for the precise phenotyping and exploration of genetic resources in foxtail millet, potentially enhancing its utilisation.
Why it matches plant phenotyping methodsフォックステールミレットの生育形質を画像から抽出する高スループット画像システムと画像処理手法を開発・評価しており、表現型取得法が研究の中心である。
abstractThis study introduces a high-throughput imaging system (HIS) with advanced image processing techniques to enhance monitoring efficiency and data quality.
Reproduction assets foundThe paper's authors explicitly state that the source code of their foxtail millet image processing program (used to extract phenotypic i-traits such as PH, CHA, SPA, compactness indices, and colour pixel features) is publicly available on GitHub. The MDSi database URL is a general transcriptomic resource, not a paper-Code · publicmation about plant health and physiological responses.
The program was created using the OpenCV library within the Microsoft Visual Studio (C++) environment. The extracted phenotypic data were converted to CSV file format for further analysis and storage. The source code of the program is available for download and reference at https://github.com/ScreenPlant/Foxtail-Millet-Image-Processing .
Analysis of i-traits throughout the entire growth periodOpen asset ↗ScreenPlant/Foxtail-Millet-Image-Processinglines:50-61Code / dataset availability confirmedOpenAlex · Europe PMC · bioRxiv · checked 15 Sept 2026
Abstract Background Automating stomatal trait measurement has gained popularity because of their inherent importance for field phenotyping application as stomata are critical for both carbon capture and water use efficiency in plants. Such tool has been reported for rice, wheat, tomato, barley and oil palm. However, none exist yet for canola, which is an important economic and agronomic crop globally. Results We developed a new toolkit called Stomatal Comprehensive Automated Neural Network or SCAN by combining the use of high-resolution portable digital microscopy with machine learning based on You Only Look Once algorithm (YOLOv8). Digital micrographs of leaf surfaces enter the SCAN pipeline, which includes stomata detection, stomata segmentation and stomatal pore segmentation models, to measure stomatal density, stomatal size and stomatal pore area, respectively. In addition to SCAN’s ability to measure leaf stomatal traits in canola at 89 to 94% accuracy, we also showed that SCAN can be used to predict stomatal density even in species not included in the training set such as Arabidopsis, tobacco, rice, wheat, maize and proso millet. SCAN was designed for the biological science community with the premise that users are not required to possess advanced programming capabilities to manage dependency prerequisites, execute the models, and integrate the analysis. This was achieved by packaging the models into a desktop application system that can be accessed offline. Conclusion Overall, SCAN provides a non-destructive, real-time, portable, and high-throughput measurement of leaf stomatal traits in canola. The minimised hardware requirement and user-friendly desktop application system make SCAN suitable for field phenotyping application.
Why it matches plant phenotyping methodsカノーラ葉の気孔形質を画像と機械学習で自動抽出するツールを開発し、精度検証と他種での適用性評価を行っており、フェノタイピング手法が中心である。
abstractWe developed a new toolkit called Stomatal Comprehensive Automated Neural Network or SCAN by combining the use of high-resolution portable digital microscopy with machine learning based on You Only Look Once algorithm (YOLOv8).
Reproduction assets foundThe paper's authors publicly deposit the SCAN pipeline's model weights, hyperparameters, training scripts, and datasets in the FD_detection GitHub repository, and provide the SCAN application itself (with download and demonstration) in a second GitHub repository. Both are paper-specific, public, and actionable.Code · publicin Table S1.
123
124
The training tasks were carried out on an Ubuntu 20.04 Linux server at the Research School of Biology in
125
Australian National University, using two Nvidia A30 (24G) Graphic Processing Units (GPUs). The full
126
details of models’ weights, hyperparameters, training scripts and datasets can be found at
127
https://github.com/William-Yao0993/FD_detection.128
129
Model evaluation
130
131
Mean Average Precision (mAP, Fig. 3) and F1 score were used to assess model ability (Fig. 4). mAP is
132
calculated as the mean value of each class area under the precision-recall curve over thresholds, and the
133
F1 score is the harmonic mean of precision and recall. The formulas are deOpen asset ↗William-Yao0993/FD_detectionpdf-raw-page:4 lines:1-81Code / dataset availability confirmedCrossref · checked 14 Sept 2026
Abstract The potential of single date fully Polarimetric RADARSAT‐2 data in retrieving crop biophysical parameters using Machine Learning techniques was investigated. Various polarimetric parameters along with coherent and incoherent decomposition techniques were assessed for its sensitivity toward crop parameters like Wet and Dry Biomass, Crop Height, Leaf Area Index and Vegetation Water Content. A set of 39 polarimetric observables extracted from the Quad‐Pol data were used for regression analysis. In this study two Machine Learning techniques Random Forest Regression (RFR) and Multiple Linear Regression (MLR) models were assessed for the prediction of Wet Biomass (gm −2 ) and Height (cm). The most significant (6 out of 39) variables were applied for prediction. The results revealed that RFR algorithm performed better than MLR. The coefficient of determination ( R 2 ) and root‐mean‐square‐error of estimating wet biomass and height were 0.646, 655.65 (gm −2 ) and 0.71, 14.5 (cm) respectively in RFR and 0.566, 683.86 (gm −2 ) and 0.65, 16.14 (cm) respectively in MLR. Thus this study explored the effective application of quad‐pol data for assessing sensitivity and accurate retrieval of parameters using optimum PolSAR observables.
Why it matches plant phenotyping methodsPolSARセンサーデータと機械学習により、作物のバイオマスおよび草丈を推定し、回帰手法の性能を比較・評価しているため、表現型取得手法が中心である。
abstractThe potential of single date fully Polarimetric RADARSAT‐2 data in retrieving crop biophysical parameters using Machine Learning techniques was investigated.
Reproduction assets foundThe paper's Data Availability Statement deposits two paper-specific public assets: the ground-truth field campaign dataset (GT Points, Zenodo 10.5281/zenodo.10403380) and the authors' RFR/MLR prediction code (Zenodo 10.5281/zenodo.10403352). Generic ESA software (PolSARpro, SNAP) is excluded as a general library.Dataset · publicEarth and Space Science
THULASIRAMAN ET AL.
10.1029/2022EA002799
17 of 20
Appendix A: Supplementary Data
Supplementary data to this article, ground truth points collected during 12 August 2019, campaign is provided in
.xlsx format (https://doi.org/10.5281/zenodo.10403380).Appendix B: RFR and MLR Algorithm
The algorithm applied for Random Forest and MLR is displayed below in Figure B1.
Data Availability Statement
The fully polarimetric RADARSAT-2 data was purchased from MDA corporation. The field data collected
during the study can be accessed from Supporting Information section (Appendix A) (htOpen asset ↗Zenodo · 10.5281/zenodo.10403380pdf-raw-page:17 lines:1-77Code / dataset availability confirmedOpenAlex · Europe PMC · Crossref · checked 8 Sept 2026
Abstract Background: A tiller is a branch on a grass plant, and the number of tillers is one of the most important determinants of yield. Traditionally, the tiller number is usually counted by hand, and so an automated approach is necessary for high-throughput phenotyping. Conventional methods use heuristic features to estimate the tiller number. Based on the successful application of DNNs in the field of computer vision, the use of DNN-based features instead of heuristic features is expected to improve the estimation accuracy. However, as DNNs generally require large volumes of data for training, it is difficult to apply them to estimation problems for which large training datasets are unavailable. In this paper, we use two strategies to overcome the problem of insufficient training data: the use of a pretrained DNN model and the use of pretext tasks for learning the feature representation. We extract features using the resulting DNNs and estimate the tiller numbers through a regression technique. Results: We conducted experiments using a dataset of Setaria viridis. Experiments show that the proposed methods using a pretrained model and specific pretext tasks achieve better performance than the conventional method. The best mean absolute error between the hand-labeled and estimated tiller numbers by the proposed method is 0.57. Conclusions: We realized applying DNN methods to tiller number estimation methods by using pretext tasks. The proposed method outperformed the conventional approach.
Why it matches plant phenotyping methods深層ニューラルネットワークを用いて植物の分げつ数を自動推定する手法を開発・比較しており、植物表現型の取得が研究の中心である。
abstractan automated approach is necessary for high-throughput phenotyping
Reproduction assets foundThe paper's tiller estimation experiments use the Setaria viridis image dataset (25,570 images, 576 with hand-labeled tiller numbers), which the authors explicitly state is publicly available in the figshare repository. No author analysis code or trained models are reported as deposited.Dataset · publicThe dataset analyzed during the current study are available in the figshare repository,
https://figshare.com/articles/dataset/DDPSC_Phenotyping_Manuscript_1_Files/1272859 [11].Open asset ↗figshare · 1272859pdf-page:12 lines:1-70Code / dataset availability confirmedEurope PMC · checked 14 Sept 2026
Achieving global goals on sustainable nutrition, health, and wellbeing will depend on delivering enhanced diets to humankind. This will require, among others, instantaneous access to information on food quality at key points within agri-food systems. Although stationary methods are usually used to quantify grain quality (wet-lab chemistry, benchtop NIR spectrometer); these do not suit many required user-cases, such as stakeholders in decentralized agri-food-chains that are typical for emerging economies. Therefore, we explored new technologies and models that might aid these particular user-cases. For this purpose, we generated the NIR spectra of 328 grain samples from multiple cereals (finger millet, foxtail millet, maize, pearl millet, sorghum) with a standard benchtop NIR Spectrometer (DS2500, FOSS) and a novel mobile NIR-based sensor (HL-EVT5, Hone). We explored a range of classical deterministic and novel machine learning (ML)-driven models to build calibrations out of the NIR spectra. We were able to build relevant calibrations out of both types of spectra. At the same time, ML-based methods enhanced the prediction capacity of calibration models compared to classical deterministic methods. We also documented that the prediction of grain protein content based on NIR spectra generated by a mobile sensor (HL-EVT5, Hone) was highly relevant for quantitative protein predictions (R2 = 0.91, RMSE = 0.97, RPD = 3.48). Thus, the findings of this study lay the foundations on which to expand the utilization of NIR spectroscopy applications for agricultural research and development.
Why it matches plant phenotyping methods穀粒という植物器官のタンパク質含量をNIRセンサーと機械学習で推定する校正モデルを開発・評価しており、形質取得・抽出法が研究の中心である。
abstractWe explored a range of classical deterministic and novel machine learning (ML)-driven models to build calibrations out of the NIR spectra.
Reproduction assets foundThe authors explicitly state that the custom CNN analysis code for this paper's NIR protein-prediction models is publicly available on GitHub at the authors' repository URL, which matches an allowed URL. Supplementary tables are only referenced via a placeholder (www.mdpi.com/xxx/s1) and are not actionable; the Video SCode · publicced by the quality
and size of the datasets used for training the model. To minimize the “over-fitting” error,
the large dataset was used and split carefully to include the different multi-cereal species
in both the calibration and validation dataset (as described in section 2.5.1). The code is
available on the GitHub platform (https://github.com/adamavip/nirs-protein-prediction)
and its particular parts can be now utilized to enhance and develop other pipelines and
products.
For our dataset, the algorithms built using ML-based methods (particularly, the
stacked ensemble model via Hone Create and custom-designed CNN; section 3.4)
achieved the better comparative metrics for both spectra typesOpen asset ↗adamavip/nirs-protein-predictionpdf-raw-page:14 lines:1-54Code / dataset availability confirmedEurope PMC · checked 9 Sept 2026
Decrease in crop yield and degradation in product quality due to plant diseases such as rust and blast in pearl millet is the cause of concern for farmers and the agriculture industry. The stipulation of expert advice for disease identification is also a challenge for the farmers. The traditional techniques adopted for plant disease detection require more human intervention, are unhandy for farmers, and have a high cost of deployment, operation, and maintenance. Therefore, there is a requirement for automating plant disease detection and classification. Deep learning and IoT-based solutions are proposed in the literature for plant disease detection and classification. However, there is a huge scope to develop low-cost systems by integrating these techniques for data collection, feature visualization, and disease detection. This research aims to develop the 'Automatic and Intelligent Data Collector and Classifier' framework by integrating IoT and deep learning. The framework automatically collects the imagery and parametric data from the pearl millet farmland at ICAR, Mysore, India. It automatically sends the collected data to the cloud server and the Raspberry Pi. The 'Custom-Net' model designed as a part of this research is deployed on the cloud server. It collaborates with the Raspberry Pi to precisely predict the blast and rust diseases in pearl millet. Moreover, the Grad-CAM is employed to visualize the features extracted by the 'Custom-Net'. Furthermore, the impact of transfer learning on the 'Custom-Net' and state-of-the-art models viz. Inception ResNet-V2, Inception-V3, ResNet-50, VGG-16, and VGG-19 is shown in this manuscript. Based on the experimental results, and features visualization by Grad-CAM, it is observed that the 'Custom-Net' extracts the relevant features and the transfer learning improves the extraction of relevant features. Additionally, the 'Custom-Net' model reports a classification accuracy of 98.78% that is equivalent to state-of-the-art models viz. Inception ResNet-V2, Inception-V3, ResNet-50, VGG-16, and VGG-19. Although the classification of 'Custom-Net' is comparable to state-of-the-art models, it is effective in reducing the training time by 86.67%. It makes the model more suitable for automating disease detection. This proves that the proposed model is effective in providing a low-cost and handy tool for farmers to improve crop yield and product quality.
Why it matches plant phenotyping methods画像とIoTデータから作物病害を推定する低コストの収集・分類フレームワークを開発し、モデル性能と特徴抽出を評価しており、植物病態の表現型取得が中心である。
abstractTherefore, there is a requirement for automating plant disease detection and classification.
Reproduction assets foundThe authors explicitly state that the pearl millet blast/rust image dataset collected as part of this research is publicly available on Kaggle, making it a paper-specific, publicly actionable phenotype image dataset. No author analysis code or trained model checkpoints are reported as publicly available.Dataset · publicData prepared as a part of this research is available at https://www.kaggle.com/kalpitgupta/blast-and-rust-compressed . The researchers who wish to use the dataset available at the above link must cite this article.Open asset ↗Kaggle · kalpitgupta/blast-and-rust-compressedlines:417-422Code / dataset availability confirmedEurope PMC · checked 9 Sept 2026
Background Assessment of seed germination is an essential task for seed researchers to measure the quality and performance of seeds. Usually, seed assessments are done manually, which is a cumbersome, time consuming and error-prone process. Classical image analyses methods are not well suited for large-scale germination experiments, because they often rely on manual adjustments of color-based thresholds. We here propose a machine learning approach using modern artificial neural networks with region proposals for accurate seed germination detection and high-throughput seed germination experiments. Results We generated labeled imaging data of the germination process of more than 2400 seeds for three different crops, Zea mays (maize), Secale cereale (rye) and Pennisetum glaucum (pearl millet), with a total of more than 23,000 images. Different state-of-the-art convolutional neural network (CNN) architectures with region proposals have been trained using transfer learning to automatically identify seeds within petri dishes and to predict whether the seeds germinated or not. Our proposed models achieved a high mean average precision (mAP) on a hold-out test data set of approximately 97.9%, 94.2% and 94.3% for Zea mays, Secale cereale and Pennisetum glaucum respectively. Further, various single-value germination indices, such as Mean Germination Time and Germination Uncertainty, can be computed more accurately with the predictions of our proposed model compared to manual countings. Conclusion Our proposed machine learning-based method can help to speed up the assessment of seed germination experiments for different seed cultivars. It has lower error rates and a higher performance compared to conventional and manual methods, leading to more accurate germination indices and quality assessments of seeds.
Why it matches plant phenotyping methods種子の発芽状態を画像と機械学習で自動抽出し、手動計数と性能比較しているため、植物表現型取得法が中心です。
abstractWe here propose a machine learning approach using modern artificial neural networks with region proposals for accurate seed germination detection and high-throughput seed germination experiments.
Reproduction assets foundThe paper's authors publicly released the labeled germination image dataset (~24,000 annotated images of 2449 seeds) on Mendeley Data and their machine learning analysis code on GitHub, both explicitly stated in the Availability of data and materials section.Dataset · publicThe generated and labeled training data is freely available on Mendeley Data: http://dx.doi.org/10.17632/4wkt6thgp6.2 .Open asset ↗Mendeley Data · 10.17632/4wkt6thgp6.2lines:164-248Code · publicThe code for our proposed machine learning–based model can be found on GitHub: https://github.com/grimmlab/GerminationPrediction .Open asset ↗GitHub · grimmlab/GerminationPredictionlines:164-248Code / dataset availability confirmedEurope PMC · OpenAlex · checked 14 Sept 2026
Genetic selection for whole-plant water use efficiency (yield per transpiration; WUE plant ) in any crop-breeding programme requires high-throughput phenotyping of component traits of WUE plant such as intrinsic water use efficiency (WUE i ; CO 2 assimilation rate per stomatal conductance). Measuring WUE i by gas exchange measurements is laborious and time consuming and may not reflect an integrated WUE i over the life of the leaf. Alternatively, leaf carbon stable isotope composition (δ 13 C leaf ) has been suggested as a potential time-integrated proxy for WUE i that may provide a tool to screen for WUE plant . However, a genetic link between δ 13 C leaf and WUE plant in a C 4 species has not been well established. Therefore, to determine if there is a genetic relationship in a C 4 plant between δ 13 C leaf and WUE plant under well watered and water-limited growth conditions, a high-throughput phenotyping facility was used to measure WUE plant in a recombinant inbred line (RIL) population created between the C 4 grasses Setaria viridis and S. italica. Three quantitative trait loci (QTL) for δ 13 C leaf were found and co-localized with transpiration, biomass accumulation, and WUE plant . Additionally, WUE plant for each of the δ 13 C leaf QTL allele classes was negatively correlated with δ 13 C leaf , as would be predicted when WUE i influences WUE plant . These results demonstrate that δ 13 C leaf is genetically linked to WUE plant , likely to be through their relationship with WUE i , and can be used as a high-throughput proxy to screen for WUE plant in these C 4 species.
Why it matches plant phenotyping methods葉の炭素安定同位体比を全植物体の水利用効率の高スループット代理指標として検証し、WUEとの遺伝的関連を評価しており、表現型取得・スクリーニング法が中心的です。
abstractleaf carbon stable isotope composition (δ 13 C leaf ) has been suggested as a potential time-integrated proxy for WUE i that may provide a tool to screen for WUE plant
Reproduction assets foundThe paper's QTL analysis was performed with the authors' custom Python and R scripts, publicly deposited on GitHub (foxy_qtl_pipeline). No public phenotype dataset URL is stated in the supplied blocks.Code · publicin Feldman et al. (2017); Feldman et al. (2018) were
308
used in this study, and here the methods used are repeated. QTL mapping was
309
performed on day 27 within each treatment group using functions in the R/qtl and
310
funqtl packages (Kwak et al., 2016). The functions were called by a set of custom
311
Python and R scripts (https://github.com/maxjfeldman/foxy_qtl_pipeline). Two
312
complimentary analysis methods were utilized. First, a single QTL model genome
313
scan using Haley-Knott regression was performed to identify QTL exhibiting LOD
314
score peaks greater than a permutation-based significance threshold (α = 0.05, n =
315
1000). Second, a stepwise forward/backward selection proceOpen asset ↗maxjfeldman/foxy_qtl_pipelinepdf-raw-page:11 lines:1-77Code / dataset availability confirmedbioRxiv · OpenAlex · Europe PMC · checked 15 Sept 2026
Association mapping studies have enabled researchers to identify candidate loci for many important environmental resistance factors, including agronomically relevant resistance traits in plants. However, traditional genome-by-environment studies such as these require a phenotyping pipeline which is capable of accurately and consistently measuring stress responses, typically in an automated high-throughput context using image processing. In this work, we present Latent Space Phenotyping (LSP), a novel phenotyping method which is able to automatically detect and quantify response to treatment directly from images. Using two synthetically generated image datasets, we first show that LSP is able to successfully recover the simulated QTL in both simple and complex synthetic imagery. We then demonstrate an example application of an interspecific cross of the model C4 grass Setaria. We propose LSP as an alternative to traditional image analysis methods for phenotyping, enabling association mapping studies without the need for engineering complex image processing pipelines.
Why it matches plant phenotyping methods画像から処理応答を自動検出・定量する新規フェノタイピング手法を提案し、合成データで検証した方法開発研究。
abstractIn this work, we present Latent Space Phenotyping (LSP), a novel phenotyping method which is able to automatically detect and quantify response to treatment directly from images.
Reproduction assets foundThe paper provides two paper-specific public assets: the LSP-Lab implementation of the Latent Space Phenotyping method on GitHub, and a figshare deposit containing the full datasets and utility scripts needed to reproduce the paper's results and figures. Setaria and sorghum image datasets are third-party (Baxter group)Dataset · publicFunding
This research was funded by a Canada First Research Excellence Fund grant from the
Natural Sciences and Engineering Research Council of Canada.
Data Availability
Full datasets and utility scripts needed for reproducing the results and figures presented in
Section 3 can be found at https://figshare.com/s/f710381c04c01e2ba319. The data
for the Setaria RIL experiment and the sorghum experiment are available from the sources
referenced by the authors of these datasets [8, 34].
References
[1] Virtual laboratory. http://www.algorithmicbotany.org/virtual_laboratory/.Accessed: 2017-08-01.
[2] Georgios Arvanitidis, Lars Kai Hansen, and Søren Hauberg. LatenOpen asset ↗figshare · f710381c04c01e2ba319pdf-raw-page:17 lines:1-44Code / dataset availability confirmedEurope PMC · checked 15 Sept 2026
MilletFlowerClassificationTrackingGrowth / development / phenology
In many plant species, the time of day at which flowers open to permit pollination is tightly regulated. Proper time of flower opening, or Time of Day of Anther Appearance ( TAA ), may coordinate flowering opening with pollinator activity or may shift temperature sensitive developmental processes to cooler times of the day. The genetic mechanisms that regulate the timing of this process in cereal crops are unknown. To address this knowledge gap, it is necessary to establish a monocot model system that exhibits variation in TAA. Here, we examine the suitability of Setaria viridis , the model for C4 photosynthesis, for such a role. We developed an imaging system to monitor the temporal regulation of growth, flower opening time, and other physiological characteristics in Setaria. This system enabled us to compare Setaria varieties Ames 32254, Ames 32276, and PI 669942 variation in growth and daily flower opening time. We observed that TAA occurs primarily at night in these three Setaria accessions. However, significant variation between the accessions was observed for both the ratio of flowers that open in the day vs. night and the specific time of day where the rate is maximal. Characterizing this physiological variation is a requisite step toward uncovering the molecular mechanisms regulating TAA. Leveraging the regulation of TAA could provide researchers with a genetic tool to improve crop productivity in new environments.
Why it matches plant phenotyping methodsSetariaの花開時刻や成長を時系列画像とニューラルネットワーク・運動解析で取得するイメージングシステムを開発しており、植物形質の取得手法が研究の中心です。
titleNeural Net Classification Combined With Movement Analysis to Evaluate Setaria viridis as a Model System for Time of Day of Anther Appearance.
Reproduction assets foundThe paper's phenotyping analysis code (camera optimization loop, day-time classification function and trained model, bristle density and growth scripts) is publicly available in the authors' DohertyLab GitHub repository, and a sample A10 day/night image dataset is deposited on Harvard Dataverse. Supplementary material,Code · publiclines never intersect, then the row of the first white pixel not classified as a bristle from the top of the image was used. The bristle density was then calculated for each image in a set. All day time images after panicle emergence were averaged to determine the final bristle density for that panicle. Scripts are available at https://github.com/DohertyLab/Setaria-Flower-Opening-Time .
Quantification of growth
ImageJ Fiji ( https://imagej.net/Fiji/Downloads ) was used to analyze the daily growth pattern of seedlings from 3 to 10 days after emergence. Dawn and dusk images at 8 a.m. and 8 p.m. were imported as an image sequence. In three replicates of each variety the second and/or third leavOpen asset ↗DohertyLab/Setaria-Flower-Opening-Timelines:43-49