were calculated based on its gradient to the disease prediction, which was equal to the weights of the last fully connected layer. A ReLU function was applied to filter negative input (Figure 2b). A python package was used to visualize the Grad-CAM (https://github.com/jacobgil/pytorch-grad-cam).Image labeling Rooster software (https://github.com/12HuYang/Rooster) was used to label tile images into disease or non-disease classes by easily clicking it with a mouse. Rooster was developed with python and can split raw images into tiles (e.g., 224 × 224 pixels) by defining column and row numbers. A semi-automatic image labeling that combines machine- and human labeling was implemented in Ro
Open resource ↗12HuYang/Rooster · pdf-raw-page:18 lines:1-30Unverified paper record
Affordable High Throughput Field Detection of Wheat Stripe Rust Using Deep Learning with Semi-Automated Image Labeling
Preprints.org · 19 Apr 2022 · 10.20944/preprints202204.0177.v1
Abstract
Stripe rust (caused by Puccinia striiformis f. sp. tritici) is one of the most devastating diseases of wheat and causes large-scale epidemics and severe yield loss. Applying fungicides during early epidemic development is crucial to controlling the disease but is often challenged by resource-limited human visual scouting. Deep learning has the potential to process images and videos captured from affordable devices to empower high-throughput phenotyping for early detection of stripe rust for timely application of fungicides and improve control efficiency. Here, we developed RustNet, a neural network-based image classifier, for efficiently monitoring fields for stripe rust. RustNet was built on a ResNet-18 architecture pre-trained with ImageNet Large-Scale Visual Recognition Challenge (ILSVRC) dataset using transfer learning. RGB images and videos of multiple wheat fields with different wheat types (winter and spring wheat), conditions (irrigated and non-irrigated), and locations were acquired using smartphones or unmanned aerial vehicles near the canopy. A semi-automated image labeling approach was conducted to improve labeling efficiency by combining automated machine labeling and human correction. Cross-validations across multiple categories (sensor platforms, wheat types, and locations) achieved Area Under Curve from 0.72 to 0.87. Independent validation on a published dataset from Germany achieved accuracies ranging from 0.79 to 0.86. The visualization of the last convolutional layer of RustNet demonstrated the identification of pixels with stripe rust. RustNet is freely available at https://zzlab.net/RustNet.
Plant phenotyping relevance
小麦のストライプさび病という植物状態を画像から検出する深層学習手法を開発し、複数条件で交差検証・独立検証しており、表現型取得手法が研究の中心です。
abstractDeep learning has the potential to process images and videos captured from affordable devices to empower high-throughput phenotyping for early detection of stripe rust
abstractHere, we developed RustNet, a neural network-based image classifier, for efficiently monitoring fields for stripe rust.
abstractCross-validations across multiple categories (sensor platforms, wheat types, and locations) achieved Area Under Curve from 0.72 to 0.87. Independent validation on a published dataset from Germany achieved accuracies ranging from 0.79 to 0.86.
Code and data availability
The authors publicly released the RustNet trained model (integrated into Rooster) and the Rooster semi-automated image-labeling software used to produce this paper's wheat stripe rust phenotyping analysis, with explicit availability statements and URLs.
This is an automatically classified, unverified record. Curator approval is required before any resource enters the Catalog.