Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability
Abstract Convolutional Neural Networks (CNNs) have proven effective in automated plant disease diagnosis, significantly contributing to crop health monitoring. However, their limited interpretability hinders practical deployment in real-world agricultural settings. To address this, we explore advanc...
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Springer
2025-05-01
|
| Series: | International Journal of Computational Intelligence Systems |
| Subjects: | |
| Online Access: | https://doi.org/10.1007/s44196-025-00835-2 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1849730851864051712 |
|---|---|
| author | Alfonso González-Briones Sebastián López Florez Pablo Chamoso Luis F. Castillo-Ossa Emilio S. Corchado |
| author_facet | Alfonso González-Briones Sebastián López Florez Pablo Chamoso Luis F. Castillo-Ossa Emilio S. Corchado |
| author_sort | Alfonso González-Briones |
| collection | DOAJ |
| description | Abstract Convolutional Neural Networks (CNNs) have proven effective in automated plant disease diagnosis, significantly contributing to crop health monitoring. However, their limited interpretability hinders practical deployment in real-world agricultural settings. To address this, we explore advanced CNN architectures, namely ResNet-50 and EfficientNet, augmented with attention mechanisms. These models enhance accuracy by optimizing depth, width, and resolution, while attention layers improve transparency by focusing on disease-relevant regions. Experiments using the PlantVillage dataset show that basic CNNs achieve 46.69% accuracy, while ResNet-50 and EfficientNet attain 63.79% and 98.27%, respectively. On a 39-class extended dataset, our proposed EfficientNet-B0 with attention (EfficientNetB0-Attn), integrating an attention module at layer 262, achieves 99.39% accuracy. This approach significantly enhances interpretability without compromising performance. The attention module generates weights via backpropagation, allowing the model to emphasize disease-relevant image regions, thereby enhancing both accuracy and interpretability. |
| format | Article |
| id | doaj-art-c00c87d8501e46d58bcd5d901f45278c |
| institution | DOAJ |
| issn | 1875-6883 |
| language | English |
| publishDate | 2025-05-01 |
| publisher | Springer |
| record_format | Article |
| series | International Journal of Computational Intelligence Systems |
| spelling | doaj-art-c00c87d8501e46d58bcd5d901f45278c2025-08-20T03:08:44ZengSpringerInternational Journal of Computational Intelligence Systems1875-68832025-05-0118111510.1007/s44196-025-00835-2Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and InterpretabilityAlfonso González-Briones0Sebastián López Florez1Pablo Chamoso2Luis F. Castillo-Ossa3Emilio S. Corchado4Grupo de Investigación BISITE, Departamento de Informática y Automática, Facultad de Ciencias, University of Salamanca, Instituto de Investigación Biomédica de SalamancaGrupo de Investigación BISITE, Departamento de Informática y Automática, Facultad de Ciencias, University of Salamanca, Instituto de Investigación Biomédica de SalamancaGrupo de Investigación BISITE, Departamento de Informática y Automática, Facultad de Ciencias, University of Salamanca, Instituto de Investigación Biomédica de SalamancaUniversidad de CaldasGrupo de Investigación BISITE, Departamento de Informática y Automática, Facultad de Ciencias, University of Salamanca, Instituto de Investigación Biomédica de SalamancaAbstract Convolutional Neural Networks (CNNs) have proven effective in automated plant disease diagnosis, significantly contributing to crop health monitoring. However, their limited interpretability hinders practical deployment in real-world agricultural settings. To address this, we explore advanced CNN architectures, namely ResNet-50 and EfficientNet, augmented with attention mechanisms. These models enhance accuracy by optimizing depth, width, and resolution, while attention layers improve transparency by focusing on disease-relevant regions. Experiments using the PlantVillage dataset show that basic CNNs achieve 46.69% accuracy, while ResNet-50 and EfficientNet attain 63.79% and 98.27%, respectively. On a 39-class extended dataset, our proposed EfficientNet-B0 with attention (EfficientNetB0-Attn), integrating an attention module at layer 262, achieves 99.39% accuracy. This approach significantly enhances interpretability without compromising performance. The attention module generates weights via backpropagation, allowing the model to emphasize disease-relevant image regions, thereby enhancing both accuracy and interpretability.https://doi.org/10.1007/s44196-025-00835-2Convolutional neural network (CNN )EfficientNetResNet-50EXplainable artificial intelligence (XAI) |
| spellingShingle | Alfonso González-Briones Sebastián López Florez Pablo Chamoso Luis F. Castillo-Ossa Emilio S. Corchado Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability International Journal of Computational Intelligence Systems Convolutional neural network (CNN ) EfficientNet ResNet-50 EXplainable artificial intelligence (XAI) |
| title | Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability |
| title_full | Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability |
| title_fullStr | Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability |
| title_full_unstemmed | Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability |
| title_short | Enhancing Plant Disease Detection: Incorporating Advanced CNN Architectures for Better Accuracy and Interpretability |
| title_sort | enhancing plant disease detection incorporating advanced cnn architectures for better accuracy and interpretability |
| topic | Convolutional neural network (CNN ) EfficientNet ResNet-50 EXplainable artificial intelligence (XAI) |
| url | https://doi.org/10.1007/s44196-025-00835-2 |
| work_keys_str_mv | AT alfonsogonzalezbriones enhancingplantdiseasedetectionincorporatingadvancedcnnarchitecturesforbetteraccuracyandinterpretability AT sebastianlopezflorez enhancingplantdiseasedetectionincorporatingadvancedcnnarchitecturesforbetteraccuracyandinterpretability AT pablochamoso enhancingplantdiseasedetectionincorporatingadvancedcnnarchitecturesforbetteraccuracyandinterpretability AT luisfcastilloossa enhancingplantdiseasedetectionincorporatingadvancedcnnarchitecturesforbetteraccuracyandinterpretability AT emilioscorchado enhancingplantdiseasedetectionincorporatingadvancedcnnarchitecturesforbetteraccuracyandinterpretability |