An Optimized CNN Model for Accurate and Fast Apple Leaf Disease Classification

Millions of people worldwide rely on apple cultivation for their livelihood, making its productivity vital for both food security and economic stability. However, apple crops are frequently affected by various leaf diseases, which can significantly hinder growth and reduce yield. Accurate and early identification of these diseases is therefore critical. This research presents a lightweight Convolutional Neural Network (CNN)-based framework for the multi-class classification of apple leaf diseases, targeting four distinct classes. The model is trained and evaluated on a subset of the PlantVillage dataset containing 3,171 labeled images of apple leaves with different disease conditions. The proposed CNN architecture integrates batch normalization, dropout, and global average pooling to enhance learning efficiency, reduce overfitting, and improve feature representation. Achieving a testing accuracy of 98%, the model demonstrates strong generalization capability while maintaining low computational complexity, making it highly suitable for real-time disease detection in agricultural settings.