CottonLeafVision: An Explainable and Robust Deep Learning Framework for Cotton Leaf Disease Classification

Researchers have developed CottonLeafVision, an explainable and robust deep learning framework for classifying cotton leaf diseases, achieving high accuracy and enhancing model reliability through techniques like Gradient-weighted Class Activation Mapping (Grad-CAM) ^[1].

Cotton is a highly economically beneficial crop globally, and precise identification of cotton leaf diseases is crucial for economic stability. The CottonLeafVision framework uses deep learning to classify cotton leaf diseases accurately, with DenseNet201 achieving a classification accuracy of 98% on a publicly available cotton leaf disease image dataset ^[1]. To enhance model reliability and interpretability, the framework implements techniques such as Grad-CAM and adversarial training to increase noise resistance. Adversarial training was also used in another study on a lemon leaf disease dataset, which contained 1354 images across 9 classes, split into training, testing, and validation sets after preprocessing ^[2]. The development of such frameworks highlights the potential of deep learning in real-life agriculture disease management situations.