Automated Diagnosis of Tomato Leaf Diseases with Attention-Enhanced Feature Aggregation

Abstract

Tomato crop yield can be enhanced by using advanced agricultural technologies when plant leaf diseases are detected early. This article proposes a new tomato disease classification model, called Multiscale Parallel Feature Aggregation Network with Attention Fusion (MPFAN-AF), that classifies diseases from leaf images. This model comprises parallel convolutional branches that extract multiscale features to obtain rich data, which are then fused through an attention-based fusion module that performs global average pooling and weights the channels. This model enables the network to optimize patterns associated with the disease while disregarding the diseased area’s background noise. Generalization performance is improved with dropout and L2 weight decay. The refined features are passed through a lightweight multi-layer perceptron for classification. Evaluated on a benchmark tomato leaf disease dataset, MPFAN-AF outperforms conventional Convolutional Neural Networks (CNNs) and existing attention-based models across accuracy, precision, recall, and F1-score. Overall, MPFAN-AF delivers an efficient, accurate, and interpretable solution for automated disease diagnosis in precision agriculture.

Keywords

• Tomato leaf disease classification
• Deep convolutional neural network (DCNN)
• Agricultural AI
• Multiscale feature extraction
• Attention fusion

Ethical Statement

It is declared that during the preparation process of this study, scientific and ethical principles were followed, and all the studies benefited from are stated in the bibliography.

References

1. Food and Agriculture Organization of the United Nations, “Agricultural production statistics 2000–2022 [Data set],” FAO, Rome, Italy, 2024. [Online]. Available: https://openknowledge.fao.org/bitstreams/fba4ef43-422c-4d73-886e-3016ff47df52/download
2. M. A. Ansari, S. Ahmad, and A. Mewada, “Mitigating risk in medical AI: Balancing X-ray datasets for reliable detection,” Life Cycle Reliab. Saf. Eng., pp. 1–14, 2025.
3. J. S. Prashanth, G. B. Krishna, A. V. Prasad, and P. R. Rao, “Smart farming revolution: A cutting-edge review of deep learning and IoT innovations in agriculture,” Oper. Res. Forum, vol. 6, no. 1, pp. 1–39, 2025.
4. A. Upadhyay et al., “Deep learning and computer vision in plant disease detection: A comprehensive review of techniques, models, and trends in precision agriculture,” Artif. Intell. Rev., vol. 58, no. 3, Art. no. 92, 2025.
5. M. A. Ansari, S. Upadhyay, A. Mewada, and S. Ahmad, “MPFAN: Multi-scale parallel feature aggregation network for plant disease diagnosis,” in Proc. 2025 Int. Conf. Eng. Innov. Technol. (ICoEIT), 2025, pp. 1478–1483.
6. A. Yadav, J. Ramaprabha, and G. K. Sandhia, “Tomato leaf disease detection using convolution neural network and VGG19,” in AIP Conf. Proc., vol. 3075, no. 1, 2024, Art. no. 020096.
7. M. Nawaz, T. Nazir, A. Javed, and M. Masood, “A robust deep learning approach for tomato plant leaf disease localization and classification,” Sci. Rep., vol. 12, Art. no. 18568, 2022.
8. A. Bhujel, N.-E. Kim, E. Arulmozhi, J. K. Basak, and H.-T. Kim, “A lightweight attention-based convolutional neural networks for tomato leaf disease classification,” Agriculture, vol. 12, no. 2, Art. no. 228, 2022.

9. A. T. Le, M. Shakiba, I. Ardekani, and W. H. Abdulla, “Optimizing plant disease classification with hybrid convolutional neural network–recurrent neural network and liquid time-constant network,” Appl. Sci., vol. 14, no. 19, Art. no. 9118, 2024.
10. M. Khan, F. Gulan, M. Arshad, and A. Zaman, “Early and late blight disease identification in tomato plants using a neural network-based model to augment agricultural productivity,” Sci. Prog., vol. 107, no. 3, Art. no. 00368504241275371, 2024.
11. J. Liang and W. Jiang, “A ResNet50-DPA model for tomato leaf disease identification,” Front. Plant Sci., vol. 14, Art. no. 1258658, 2023.
12. P. Bedi and P. Gole, “Plant disease detection using hybrid model based on convolutional autoencoder and convolutional neural network,” Artif. Intell. Agric., vol. 5, pp. 90–101, 2021.
13. M. Xu, S. Yoon, Y. Jeong, and D. S. Park, “Transfer learning for versatile plant disease recognition with limited data,” Front. Plant Sci., vol. 13, Art. no. 1010981, 2022.
14. A. Abade, P. A. Ferreira, and F. B. Vidal, “Plant diseases recognition on images using convolutional neural networks: A systematic review,” Comput. Electron. Agric., vol. 185, Art. no. 106125, 2021.
15. I. Pacal et al., “A systematic review of deep learning techniques for plant diseases,” Artif. Intell. Rev., vol. 57, Art. no. 304, 2024, doi: 10.1007/s10462-024-10944-7. [Online]. Available: https://doi.org/10.1007/s10462-024-10944-7
16. A. Mewada, S. K. Maurya, and M. A. Ansari, “Comparative study of artificial intelligence approaches in deceptive opinion detection,” in Proc. 2nd Int. Conf. Comput. Intell., Commun. Technol. Netw. (CICTN), 2025, pp. 999–1004.
17. S. Sharma and M. Vardhan, “MTJNet: Multi-task joint learning network for advancing medicinal plant and leaf classification,” Knowl.-Based Syst., vol. 299, Art. no. 112147, 2024.
18. S. S. Chouhan, U. P. Singh, and S. Jain, “Performance evaluation of different deep learning models used for the purpose of healthy and diseased leaves classification of Cherimoya (Annona cherimola) plant,” Neural Comput. Appl., vol. 37, no. 6, pp. 4531–4544, 2025.
19. S. Mumtaz, S. Algamdi, H. F. Alhasson, D. A. AlHammadi, A. Jalal, and H. Liu, “Leaf classification for sustainable agriculture and in-depth species analysis,” IEEE Access, 2025.
20. V. Solapure, SmartAgroTech DY, and A. Jawale, “Tomato leaf disease dataset,” Mendeley Data, V1, 2024, doi: 10.17632/zfv4jj7855.1. [Online]. Available: https://doi.org/10.17632/zfv4jj7855.1
21. A. Sharma and R. Singh, “ConvST-LSTM-Net: Convolutional spatiotemporal LSTM networks for skeleton-based human action recognition,” Int. J. Multimedia Inf. Retrieval, vol. 12, no. 2, Art. no. 34, 2023.
22. M. Shah and N. Sureja, “A comprehensive review of bias in deep learning models: Methods, impacts, and future directions,” Arch. Comput. Methods Eng., vol. 32, no. 1, pp. 255–267, 2025.