Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{195364,
        author = {Dhammapal Y. Tayade and Prof.D.N.Besekar},
        title = {Explainable Deep Learning for Cotton Leaf Disease Classification Using ResNet50 and Grad-CAM},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {10},
        pages = {7700-7712},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=195364},
        abstract = {Cotton is one of India’s most important commercial crops, yet its yield is significantly affected by foliar diseases such as Curl Virus, Fusarium Wilt, Bacterial Blight, Powdery Mildew, Target Spot, and pest-induced damage. Early identification is essential for controlling disease spread, but traditional manual diagnosis performed by farmers is slow, expertise-dependent, and often inaccurate. To address this challenge, this study proposes an explainable deep learning framework for the automatic classification of cotton leaf diseases using a ResNet50-based Convolutional Neural Network (CNN) integrated with Gradient-weighted Class Activation Mapping (Grad-CAM). A merged dataset of 5,561 cotton leaf images collected from the Akola and Yavatmal districts and supplemented with publicly available samples was used for experimentation. The images were categorized into seven classes: Bacterial Blight, Curl Virus, Fusarium Wilt, Powdery Mildew, Target Spot, Pest Damage, and Healthy leaves.
A two-stage transfer learning strategy was adopted, involving frozen feature extraction followed by selective fine-tuning of deeper layers for disease-specific representation learning. The proposed ResNet50 model achieved a test accuracy of 99.2%, demonstrating strong generalization performance.
The core contribution of this research lies in integrating Grad-CAM to provide visual explanations for model predictions. The generated heatmaps highlight disease-relevant regions such as lesions, vein distortions, fungal growth, and pest damage, thereby improving transparency and trust. The proposed framework represents a step toward Explainable AI (XAI) in agriculture and is suitable for integration into mobile or web-based decision-support systems for early disease diagnosis.},
        keywords = {Cotton leaf diseases, Deep learning, Explainable AI, Grad-CAM, CNN, ResNet50, Transfer learning.},
        month = {March},
        }