<![CDATA[Manual identification of rice leaf diseases is often inefficient. The purpose of this study is to conduct a comparative analysis of classification models for detecting four types of rice leaf diseases. The method involves feature extraction from a MobileNetV2 architecture on the data from the Rice Leaf Disease Dataset, containing 5,932 images. Four models were tested: a Prototype Classifier, MLP-Softmax, Support Vector Machine (SVM) with an RBF kernel, and a Hybrid Ensemble. The evaluation results showed that the SVM-RBF and Hybrid Ensemble models achieved the best performance with a perfect accuracy of 100%, outperforming the MLP-Softmax (99.24%) and the Prototype Classifier (79.58%). This study concludes that the synergy between MobileNetV2 features and SVM classification provides a highly accurate solution for automated rice leaf disease detection.]]>
Copyrights © 2025
