<![CDATA[Rice is of the main food commodities in Indonesia that is susceptible to various leaf diseases, one of which is Bacterial Blight, Brown Spot, and Leaf Smut. Manual identification by farmers is often less accurate and time-consuming, thus requiring a technology-based detection system. The objective of this research is to categorize rice leafdiseases through the use of deep learning with a transfer learning approach based on MobileNetV2 and MobileNetV3 architectures. The dataset, comprising 4,684 rice leaf images, was divided into training and validation sets using an 80:20 ratio. Preprocessing included resizing images to 224×224 pixels, normalization, and augmentation to increase data variation. Training was carried out across 30 epochs with a mini-batch size set to 32. while applying an EarlyStopping mechanism to reduce the likelihood of overfitting. The result of the experiment indicate that MobileNetV2 reached an 96% accuracy, while MobileNetV3 outputperformed is with an accuracy of 99%. Therefore, MobileNetV3 is more effective for rice leaf disease classification.]]>
Copyrights © 2025
