<![CDATA[Rice production in Indonesia often declines due to leaf diseases that are difficult to detect early using conventional methods. This study aims to identify rice leaf diseases based on leaf images using a Convolutional Neural Network (CNN). The dataset was obtained from an online repository (Kaggle) containing labeled images of rice leaves across several disease categories. A custom CNN model was designed and trained after applying image preprocessing (resizing to 224×224 pixels), normalization, and data augmentation to reduce overfitting. The training was conducted in the Google Colab environment using TensorFlow with train–test splits of 70:30, 80:20, and 90:10 to analyze model performance. The best result achieved a training accuracy of 83.02% and a testing accuracy of 77.33%. Furthermore, the model was compared with several widely used architectures in the literature, including ResNet50, VGG16, and EfficientNetB0. The findings indicate that the proposed custom CNN model provides competitive classification performance for early detection of rice leaf diseases and has the potential to serve as a decision-support system for farmers in rapid and efficient disease management.]]>
Copyrights © 2025
