A Review on Deep Learning Framework for Automated Classification of Banana Leaf Diseases

Banana is one of the most economically significant fruit crops worldwide, yet its productivity is severely affected by a wide range of leaf diseases that compromise yield and quality. Traditional disease identification methods rely on expert visual inspection, which is time-consuming, subjective, and often inaccessible to small-scale farmers. Recent advances in deep learning have enabled automated, accurate, and scalable solutions for banana leaf disease classification using digital images. This review presents a comprehensive analysis of deep learning frameworks developed for automated banana leaf disease detection and classification. The paper systematically examines convolutional neural networks, lightweight architectures, hybrid CNN–Transformer models, object detection frameworks, and explainable AI techniques applied to banana leaf images. Additionally, datasets, preprocessing strategies, performance evaluation metrics, and deployment considerations are discussed. The review highlights key research findings, identifies major challenges such as dataset imbalance, environmental variability, and model generalization, and outlines future research directions to enhance robustness and real-world applicability. This study aims to serve as a valuable reference for researchers and practitioners working on intelligent agricultural disease diagnosis systems.