Comparison of Convolutional Neural Network and MobileNetV2 for Potato Leaf Disease Detection

This research develops a method for detecting potato leaf diseases using Deep Learning techniques to overcome the limitations of manual methods. The research uses Convolutional Neural Network (CNN) and MobileNetV2 architectures to classify potato leaf diseases into three categories: healthy, early blight, and late blight. The dataset includes 4072 images for training, testing, and validation. Evaluation results reveal that the MobileNetV2 model achieves the highest accuracy of 93.00%, with precision, recall, and F1 scores of 93.86%, 93.33%, and 93.38%, respectively, outperforming the CNN model, which has an accuracy of 90.00%. These findings demonstrate MobileNetV2’s effectiveness in accurate disease classification, offering a viable solution for efficient detection in resource-constrained agricultural settings.