<![CDATA[Plant leaf diseases are one of the major factors that reduce agricultural productivity, making fast and accurate identification methods essential. This study aims to compare the performance of two lightweight deep learning architectures, MobileNetV3 Small and EfficientNetV2S, for plant leaf disease identification using digital images. The dataset used was the Plant_leaf_diseases_dataset_without_augmentation obtained from Mendeley Data. The research workflow consisted of image preprocessing, data augmentation, model training with hyperparameter configuration, fine-tuning, and performance evaluation using TensorFlow based on accuracy, precision, recall, F1-score, training time, model size, and inference time. The results showed that EfficientNetV2S achieved the best classification performance, with an accuracy of 97.66%, a precision of 97.80%, a recall of 97.66%, and an F1-score of 97.67%. Meanwhile, MobileNetV3 Small achieved an accuracy of 96.40%, with a model size of 9.96 MB, an inference time of 1.46 ms, and a frame rate of 686.95 FPSse findings indicate that the two models have different strengths: EfficientNetV2S excels in classification accuracy, whereas MobileNetV3 Small offers superior computational efficiency.]]>
Copyrights © 2026
