Modified CNN Architecture for Paddy Leaf Disease Detection

Paddy (rice) is the main food for the Indonesian people so that stable production is needed to meet domestic needs. However, data in 2023 showed a decline in rice production in Indonesia. The decline was due to pests and diseases that attacked rice plants. The lack of experience and knowledge of farmers in handling and detecting rice diseases is also a factor in the decline in rice productivity. This study aims to develop a computer-aided diagnosis (CAD) system for rice leaf diseases based on CNN algorithms. In creating CAD applications, this study proposes a new CNN architecture based on a modification of the MobileNetV3 architecture. The dataset used comes from direct data collection from rice fields in Indonesia which is then added with public data from Kaggle data totalling 6933 rice leaf images. Based on the results obtained, this study succeeded in surpassing the previous research benchmark with an average accuracy, precision, recall and F -1 Score value reaching 93%.