Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution
In image classification of tomato leaf diseases based on deep learning, models often focus on features such as edges, stems, backgrounds, and shadows of the experimental samples, while ignoring the features of the disease area, resulting in weak generalization ability. In this study, a self-attentio...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-09-01
|
| Series: | Horticulturae |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2311-7524/9/9/1034 |
| _version_ | 1797579818503503872 |
|---|---|
| author | Zheng Li Weijie Tao Jianlei Liu Fenghua Zhu Guangyue Du Guanggang Ji |
| author_facet | Zheng Li Weijie Tao Jianlei Liu Fenghua Zhu Guangyue Du Guanggang Ji |
| author_sort | Zheng Li |
| collection | DOAJ |
| description | In image classification of tomato leaf diseases based on deep learning, models often focus on features such as edges, stems, backgrounds, and shadows of the experimental samples, while ignoring the features of the disease area, resulting in weak generalization ability. In this study, a self-attention mechanism called GD-Attention is proposed, which considers global pixel value distribution information and guide the deep learning model to give more concern on the leaf disease area. Based on data augmentation, the proposed method inputs both the image and its pixel value distribution information to the model. The GD-Attention mechanism guides the model to extract features related to pixel value distribution information, thereby increasing attention towards the disease area. The model is trained and tested on the Plant Village (PV) dataset, and by analyzing the generated attention heatmaps, it is observed that the disease area obtains greater weight. The results achieve an accuracy of 99.97% and 27 MB parameters only. Compared to classical and state-of-the-art models, our model showcases competitive performance. As a next step, we are committed to further research and application, aiming to address real-world, complex scenarios. |
| first_indexed | 2024-03-10T22:41:18Z |
| format | Article |
| id | doaj.art-7d630a7ee4bf40a395c11dc7262f96a7 |
| institution | Directory Open Access Journal |
| issn | 2311-7524 |
| language | English |
| last_indexed | 2024-03-10T22:41:18Z |
| publishDate | 2023-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Horticulturae |
| spelling | doaj.art-7d630a7ee4bf40a395c11dc7262f96a72023-11-19T10:59:17ZengMDPI AGHorticulturae2311-75242023-09-0199103410.3390/horticulturae9091034Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value DistributionZheng Li0Weijie Tao1Jianlei Liu2Fenghua Zhu3Guangyue Du4Guanggang Ji5School of Rail Transportation, Shandong Jiaotong University, Jinan 250357, ChinaSchool of Rail Transportation, Shandong Jiaotong University, Jinan 250357, ChinaDepartment of Cyberspace Security, Qufu Normal University, Qufu 273165, ChinaInstitute of Automation, Chinese Academy of Sciences, Beijing 100049, ChinaSchool of Rail Transportation, Shandong Jiaotong University, Jinan 250357, ChinaSchool of Rail Transportation, Shandong Jiaotong University, Jinan 250357, ChinaIn image classification of tomato leaf diseases based on deep learning, models often focus on features such as edges, stems, backgrounds, and shadows of the experimental samples, while ignoring the features of the disease area, resulting in weak generalization ability. In this study, a self-attention mechanism called GD-Attention is proposed, which considers global pixel value distribution information and guide the deep learning model to give more concern on the leaf disease area. Based on data augmentation, the proposed method inputs both the image and its pixel value distribution information to the model. The GD-Attention mechanism guides the model to extract features related to pixel value distribution information, thereby increasing attention towards the disease area. The model is trained and tested on the Plant Village (PV) dataset, and by analyzing the generated attention heatmaps, it is observed that the disease area obtains greater weight. The results achieve an accuracy of 99.97% and 27 MB parameters only. Compared to classical and state-of-the-art models, our model showcases competitive performance. As a next step, we are committed to further research and application, aiming to address real-world, complex scenarios.https://www.mdpi.com/2311-7524/9/9/1034plant leaf diseaseimage recognitionattention mechanismsmart agriculture |
| spellingShingle | Zheng Li Weijie Tao Jianlei Liu Fenghua Zhu Guangyue Du Guanggang Ji Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution Horticulturae plant leaf disease image recognition attention mechanism smart agriculture |
| title | Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution |
| title_full | Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution |
| title_fullStr | Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution |
| title_full_unstemmed | Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution |
| title_short | Tomato Leaf Disease Recognition via Optimizing Deep Learning Methods Considering Global Pixel Value Distribution |
| title_sort | tomato leaf disease recognition via optimizing deep learning methods considering global pixel value distribution |
| topic | plant leaf disease image recognition attention mechanism smart agriculture |
| url | https://www.mdpi.com/2311-7524/9/9/1034 |
| work_keys_str_mv | AT zhengli tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution AT weijietao tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution AT jianleiliu tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution AT fenghuazhu tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution AT guangyuedu tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution AT guanggangji tomatoleafdiseaserecognitionviaoptimizingdeeplearningmethodsconsideringglobalpixelvaluedistribution |