Citrus Leaf Disease Detection Using GAN-Based Data Augmentation And Convolutional Neural Networks

Shubham S. Madane; Dr. Vikul J. Pawar; Dr. Kailash Kharat

doi:10.29284/25p3gb26

Authors

Shubham S. Madane Department of Computer Science and Engineering, Government College of Engineering, Aurangabad (Chhatrapati Sambhajinagar), India
Dr. Vikul J. Pawar Department of Computer Science and Engineering, Government College of Engineering, Aurangabad (Chhatrapati Sambhajinagar), India
Dr. Kailash Kharat CSMSS Chhatrapati Shahu College of Engineering, Chhatrapati Sambhajinagar, Maharashtra, India

DOI:

https://doi.org/10.29284/25p3gb26

Keywords:

Citrus disease detection, computer vision, Con- volutional Neural Network, deep learning, Generative Adversar- ial Network, ground-level farm dataset, plant pathology, precision agriculture, synthetic data augmentation.

Abstract

Citrus crops are one of the pillars of world agri- culture but, due to the presence of pathogenic diseases, the quality of the yield and the volume of production are significantly reduced. The traditional diagnostic methods that used visual examination by skilled agronomists are simply slow, subjective and operationally unfeasible at farm level. The automated deep- learning model presented in the paper aims to classify citrus leaf disease images based on a real-farm dataset of 934 images of 6 types (leaf blast (115), healthy (167), canker (163), greening (204), black spot (171), and melanose (114)). Agricultural deep learning applications are characterized by two major challenges, namely, a lack of annotated training data and severe class imbalance. Their combination resulted in disease-specific Deep Convolutional Generative Adversarial Networks (DCGANs) to create realistic diseased-leaf examples, in which 500 balanced training examples per minority class are used. A Convolutional Neural Network based on the EfficientNetB0 architecture was then fine-tuned on the synthetically enriched dataset using transfer learning. When only the original, unaugmented farm images are evaluated, it is shown that the GAN-augmented model shows a statistically sig- nificant 92.78% classification accuracy, as compared to 86.42% of the baseline achieved without synthetic augmentation. Precision and recall of the proposed method by class also confirm the strength of the offered solution regarding all types of diseases. The results provide generalizable, scalable approach to deal with data scarcity in agricultural diagnostics and provide a solution that can be implemented in the field to detect citrus diseases at an early stage.

References

[1] R. K. Patel, S. S. Shukla, and V. K. Gupta, “Citrus disease detection using deep convolutional neural networks: A comprehensive review and future directions,” IEEE Trans. AgriFood Electron., vol. 1, no. 2, pp. 112–128, 2023.

[2] M. Sharma, A. Kumar, and P. Singh, “Deep learning-based citrus greening disease detection using hyperspectral imaging and transfer learning,” Comput. Electron. Agric., vol. 208, p. 107789, 2023.

[3] F. Ahmad, A. Khan, and S. Jan, “Lightweight deep learning models for plant disease detection on resource-constrained devices: A comparative analysis,” IEEE Sensors J., vol. 23, no. 15, pp. 17234–17247, 2023.

[4] S. B. Jadhav, V. R. Udupi, and S. B. Patil, “Interpretable deep learning for plant disease classification using Grad-CAM and LIME,” IEEE Access, vol. 11, pp. 115678–115693, 2023.

[5] S. Gupta, A. Sharma, and M. Tiwari, “Conditional GAN with attention mechanism for enhanced plant disease image synthesis,” IEEE Trans. Artif. Intell., vol. 4, no. 3, pp. 567–579, 2023.

[6] J. Kim, S. Park, and H. Lee, “StyleGAN-based data augmentation for improving robustness of plant disease classification,” Comput. Electron. Agric., vol. 212, p. 108123, 2024.

[7] T. Zhang, L. Wang, and Y. Liu, “Progressive growing GAN for gen- erating diverse plant disease symptoms under varying environmental conditions,” Pattern Recognit. Lett., vol. 175, pp. 45–52, 2023.

[8] A. Kamilaris and F. X. Prenafeta-Boldu´, “Deep learning in agriculture: A survey on image-based plant disease detection,” Comput. Electron. Agric., vol. 205, p. 107612, 2023.

[9] S. Verma, A. Srivastava, and R. K. Singh, “Multi-scale convolutional neural network with attention mechanism for early plant disease detec- tion,” Comput. Electron. Agric., vol. 215, p. 108456, 2023.

[10] P. K. Sethy, S. K. Behera, and N. K. Rath, “Deep feature extraction and ensemble learning for plant disease identification in real-field conditions,” Expert Syst. Appl., vol. 235, p. 121234, 2024.

[11] L. Chen, Y. Wang, and H. Zhang, “Multi-scale attention network for citrus disease recognition in complex orchard environments,” IEEE Access, vol. 11, pp. 45231–45244, 2023.

[12] A. K. Singh, P. Kumar, and R. Singh, “GAN-based synthetic image generation for plant disease detection: A systematic review,” Eng. Appl. Artif. Intell., vol. 126, p. 107145, 2023.

[13] M. A. Rahman, M. S. Hossain, and N. Ahmed, “DCGAN with perceptual loss for high-fidelity plant disease image generation,” IEEE Access, vol. 12, pp. 15678–15692, 2024.

[14] H. Wang, J. Li, and X. Chen, “EfficientNet-based deep learning archi- tecture for plant disease classification with attention mechanisms,” IEEE Trans. AgriFood Electron., vol. 2, no. 1, pp. 34–47, 2024.

[15] S. K. Roy, M. E. Paoletti, and J. M. Haut, “UAV-based remote sensing for plant disease monitoring using deep learning: A review,” IEEE Geosci. Remote Sens. Mag., vol. 11, no. 3, pp. 82–109, 2023.

[16] V. S. Rao, K. P. Soman, and M. V. R. Varma, “Federated learning for plant disease detection in smart agriculture: Challenges and opportuni- ties,” IEEE Trans. AgriFood Electron., vol. 1, no. 3, pp. 234–249, 2023.

[17] . R. Reddy, K. P. S. Kumar, and V. N. Sastry, “Edge-AI for real- time plant disease detection: A survey of architectures, frameworks, and deployment strategies,” ACM Comput. Surv., vol. 56, no. 8, pp. 1–35, 2024.

[18] N. T. Nguyen, T. T. Le, and H. T. Nguyen, “Mobile-based plant disease detection using deep learning: A systematic review and future directions,” IEEE Internet Things J., vol. 10, no. 24, pp. 21789–21808, 2023.

[19] M. T. Islam, M. A. Hossain, and K. Ahmed, “Transformers for plant disease detection: A comprehensive review and future perspectives,” IEEE Trans. Neural Netw. Learn. Syst., vol. 35, no. 2, pp. 1456–1473, 2024.

[20] L. Deng, J. Liu, and Y. Han, “Few-shot learning for plant disease detec- tion: A comprehensive review and future challenges,” Neurocomputing, vol. 550, p. 126512, 2023.

[21] A. K. Dubey and V. Jain, “Plant disease detection using deep learning: A bibliometric analysis and future research directions,” IEEE Access, vol. 12, pp. 34567–34589, 2024.

[22] V. Kumar, A. Sharma, and P. Singh, “Comparative analysis of CNN architectures for real-time plant disease detection on edge devices,” Multimedia Tools Appl., vol. 83, no. 4, pp. 11245–11268, 2024.