Convolutional Neural Network (CNN) EfficientNet-B0 Model Architecture for Paddy Diseases Classification

Rice stands as a pivotal food commodity in Indonesia, contributing significantly to food security while witnessing a consistent rise in demand. The susceptibility of rice leaves to diseases threatens this critical role, underscoring the imperative for early and accurately automated disease identification. In this study, we employ Convolutional Neural Networks (CNN) as a deep learning methodology to classify various types of rice leaf diseases. The approach comprises distinct stages, encompassing preprocessing, modeling, and evaluation. The primary objective of this research is to achieve the classification of rice leaf diseases with both acceptable accuracy and runtime efficiency. The experimental outcomes highlight an impressive peak accuracy of 98.93% when employing data augmentation, layer normalization, and the RMSProp optimization function with a learning rate of 0.001. These results underscore the significance of diligent data preparation and model optimization in the rice leaf disease classification process.