Scholarly record
AN EFFICIENT AND INTERPRETABLE HYBRID AI FRAMEWORK FOR COCONUT LEAF DISEASE DIAGNOSIS USING CONVOLUTIONAL NEURAL NETWORKS
Abstract
Coconut cultivation is a vital component of agricultural sustainability in tropical regions, particularly in Kerala, India, where leaf diseases pose a significant threat to crop yield and plantation health. Automated diagnosis of coconut leaf diseases using vision-based deep learning models has shown promising performance, but their real-world applicability is limited by challenges such as severe class imbalance, high-dimensional feature learning, excessive computational overhead, and limited interpretability during inference. Addressing these limitations requires a mathematically grounded and computationally efficient diagnostic framework. This paper proposes an efficient and interpretable hybrid artificial intelligence framework for coconut leaf disease diagnosis that integrates probabilistic machine learning with deep convolutional neural network representations. The proposed approach formulates disease detection as a hierarchical classification problem. In the first stage, a lightweight probabilistic classifier performs binary hypothesis testing to estimate posterior class probabilities and separate healthy from diseased leaf samples, thereby reducing sample complexity and inference redundancy. In the second stage, a convolutional neural network performs multi-class disease discrimination by learning hierarchical spatial representations optimised through gradient-based cross-entropy minimisation. To enhance transparency and diagnostic reliability, gradient-based explainable artificial intelligence techniques are incorporated to generate spatial relevance maps that localise disease- affected regions and provide insight into the model’s decision process. Experimental analysis demonstrates that the proposed framework achieves competitive classification accuracy while significantly reducing inference latency and computational complexity. The results confirm the efficiency of the proposed approach for real-time, resource-constrained agricultural environments, offering a scalable, interpretable, and mathematically principled solution for automated diagnosis of coconut leaf disease.
Publication details
References13
Abade, A., Ferreira, P. A., & Vidal, F. B. (2021). Plant diseases recognition on images using convolutional neural networks: A systematic review. Computers and Electronics in Agriculture, 185, 106125. DOI: 10.1016/j.compag.2021.106125
Brar, S. K., Sharma, R., Vats, S., & Kukreja, V. (2026). A smart approach to coconut leaf spot disease classification using computer vision and deep learning technique. Proceedings of the 2026 World Conference on Communication & Computing. IEEE. DOI: 10.1109/WCONF.2023.10235251
Chawla, N. V., Bowyer, K. W., Hall, L. O., & Kegelmeyer, W. P. (2002). SMOTE: Synthetic minority over-sampling technique. Journal of Artificial Intelligence Research, 16, 321-357. DOI: 10.1613/jair.953 Coconut Development Board. (2025). Area, production and productivity of coconut in India. Ministry of Agriculture and Farmers Welfare, Government of India. https://coconutboard.gov.in/Statistics.aspx Ferentinos, K. P. (2018). Deep learning models for plant disease detection and diagnosis. Computers and Electronics in Agriculture, 145, 311-318. https://doi.org/10.1016/j.compag.2018.01.009
Gopalakrishna, K. M., & Lingaraju, R. M. (2025). A prediction of coconut and coconut leaf disease using MobileNetV2 based classification. International Journal of Electrical and Computer Engineering, 15(3), 2834-2844. DOI: 10.11591/ijece.v15i3.pp2834-2844
He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 770-778. DOI: 10.1109/CVPR.2016.90
Howard, A. G., Zhu, M., Chen, B., Kalenichenko, D., Wang, W., Weyand, T., Andreetto, M., & Adam, H. (2017). MobileNets: Efficient convolutional neural networks for mobile vision applications. arXiv. https://arxiv.org/abs/1704.04861 Kaur, A., Kukreja, V., & Garg, N. (2024). Multi-class classification of coconut leaf disease using fine-tuned ResNet50 model. Proceedings of the 2024 9th International Conference on Information, Communication and Computing Technology. IEEE. DOI: 10.1109/I2CT61223.2024.10544047
Lin, T. Y., Goyal, P., Girshick, R., He, K., & Dollar, P. (2017). Focal loss for dense object detection. Proceedings of the IEEE International Conference on Computer Vision, 2980-2988. DOI: 10.1109/ICCV.2017.324
Manoharan, S. K., Megalingam, R. K., Gopika, A., Jogesh, G., & Kunnambath, A. R. (2024). AI based early identification and severity detection of nutrient deficiencies in coconut trees. Smart Agricultural Technology, 9, 100575. DOI: 10.1016/j.atech.2024.100575
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2020). Grad-CAM: Visual explanations from deep networks via gradient-based localization. International Journal of Computer Vision, 128, 336-359. DOI: 10.1007/s11263-019-01228-7
Shorten, C., & Khoshgoftaar, T. M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6, Article 60. DOI: 10.1186/s40537-019-0197-0
Subbaian, S., Balasubramanian, A., Marimuthu, M., Chandrasekaran, S., & Muthusaravanan, G. (2024). Detection of coconut leaf diseases using enhanced deep learning techniques. Journal of Intelligent & Fuzzy Systems, 46(2), 5033-5045. DOI: 10.3233/JIFS-233831
Tan, M., & Le, Q. V. (2019). EfficientNet: Rethinking model scaling for convolutional neural networks. Proceedings of the 36th International Conference on Machine Learning, 97, 6105-6114.
Thite, S., Suryawanshi, Y., Patil, K., & Chumchu, P. (2023). Coconut (Cocos nucifera) tree disease dataset: A dataset for disease detection and classification for machine learning applications. Data in Brief, 51, 109690. DOI: 10.1016/j.dib.2023.109690