Abstract
To address the challenges of low recognition accuracy and substantial model complexity in crop disease identification models operating in complex field environments, this study proposed a novel hybrid model named ConvTransNet-S, which integrates Convolutional Neural Networks (CNNs) and transformers for crop disease identification tasks. Unlike existing hybrid approaches, ConvTransNet-S uniquely introduces three key innovations: First, a Local Perception Unit (LPU) and Lightweight Multi-Head Self-Attention (LMHSA) modules were introduced to synergistically enhance the extraction of fine-grained plant disease details and model global dependency relationships, respectively. Second, an Inverted Residual Feed-Forward Network (IRFFN) was employed to optimize the feature propagation path, thereby enhancing the model's robustness against interferences such as lighting variations and leaf occlusions. This novel combination of a LPU, LMHSA, and an IRFFN achieves a dynamic equilibrium between local texture perception and global context modeling-effectively resolving the trade-offs inherent in standalone CNNs or transformers. Finally, through a phased architecture design, efficient fusion of multi-scale disease features is achieved, which enhances feature discriminability while reducing model complexity. The experimental results indicated that ConvTransNet-S achieved a recognition accuracy of 98.85% on the PlantVillage public dataset. This model operates with only 25.14 million parameters, a computational load of 3.762 GFLOPs, and an inference time of 7.56 ms. Testing on a self-built in-field complex scene dataset comprising 10,441 images revealed that ConvTransNet-S achieved an accuracy of 88.53%, which represents improvements of 14.22%, 2.75%, and 0.34% over EfficientNetV2, Vision Transformer, and Swin Transformer, respectively. Furthermore, the ConvTransNet-S model achieved up to 14.22% higher disease recognition accuracy under complex background conditions while reducing the parameter count by 46.8%. This confirms that its unique multi-scale feature mechanism can effectively distinguish disease from background features, providing a novel technical approach for disease diagnosis in complex agricultural scenarios and demonstrating significant application value for intelligent agricultural management.
Abstract from the paper’s OpenAlex record; open copy licensed cc-by. Read at the source ↗
The text the model read, with the source record’s inline markup (italics, sub- and superscripts) rendered and its character entities decoded; quote offsets refer to the raw text.
Claims 10
Claim · model-written summary
ConvTransNet-S reached 98.85% recognition accuracy on PlantVillage.
Verbatim
e0“ConvTransNet-S achieved a recognition accuracy of 98.85% on the PlantVillage public dataset.”
t0“crop disease identification tasks”
Claim · model-written summary
ConvTransNet-S used 25.14 million parameters.
Verbatim
e0“This model operates with only 25.14 million parameters, a computational load of 3.762 GFLOPs, and an inference time of 7.56 ms.”
t0“crop disease identification tasks”
Claim · model-written summary
ConvTransNet-S had a computational load of 3.762 GFLOPs.
Verbatim
e0“a computational load of 3.762 GFLOPs”
t0“crop disease identification tasks”
Claim · model-written summary
ConvTransNet-S inference time was 7.56 ms.
Verbatim
e0“an inference time of 7.56 ms”
t0“crop disease identification tasks”
Claim · model-written summary
ConvTransNet-S reached 88.53% accuracy on field scenes.
Verbatim
e0“Testing on a self-built in-field complex scene dataset comprising 10,441 images revealed that ConvTransNet-S achieved an accuracy of 88.53%”
t0“crop disease identification tasks”
Claim · model-written summary
Field accuracy exceeded EfficientNetV2 by 14.22%.
Verbatim
e0“improvements of 14.22%, 2.75%, and 0.34% over EfficientNetV2, Vision Transformer, and Swin Transformer, respectively”
t0“crop disease identification tasks”
Claim · model-written summary
Field accuracy exceeded Vision Transformer by 2.75%.
Verbatim
e0“improvements of 14.22%, 2.75%, and 0.34% over EfficientNetV2, Vision Transformer, and Swin Transformer, respectively”
t0“crop disease identification tasks”
Claim · model-written summary
Field accuracy exceeded Swin Transformer by 0.34%.
Verbatim
e0“improvements of 14.22%, 2.75%, and 0.34% over EfficientNetV2, Vision Transformer, and Swin Transformer, respectively”
t0“crop disease identification tasks”
Claim · model-written summary
Complex-background disease accuracy was up to 14.22% higher.
Verbatim
e0“up to 14.22% higher disease recognition accuracy under complex background conditions”
t0“crop disease identification tasks”
Claim · model-written summary
Parameter count was reduced by 46.8%.
Verbatim
e0“reducing the parameter count by 46.8%”
t0“crop disease identification tasks”