Toward Reliable Tea Leaf Disease Diagnosis Using Deep Learning Model: Enhancing Robustness With Explainable AI and Adversarial Training

Researchers developed a deep learning system to automatically diagnose diseases in tea leaves, achieving 93% classification accuracy. The model uses EfficientNetB3 and DenseNet201 architectures trained on 5,278 high-resolution images of healthy and diseased leaves. It incorporates adversarial training for robustness against noisy inputs and Explainable AI (Grad-CAM) to show which parts of the leaf it's analyzing. The goal is to provide farmers with a fast, reliable tool to prevent crop loss and boost tea production quality.