Manuscript received February 21, 2026; revised April 3, 2026; accepted May 15, 2026; published October 20, 2026.

Abstract—Rice diseases pose a serious threat to agricultural productivity and food security, especially in Indonesia. This study proposes a rice disease detection system using deep learning-based rice leaf images implemented on mobile devices to provide fast, accurate, and easily accessible diagnoses. Two main architectures were used, namely You Only Look Once version 8 (YOLOv8) for real-time detection and EfficientNet for high-efficiency classification. The dataset consisted of an open-source dataset that had undergone annotation, pre-processing, and augmentation. The models implemented on mobile devices are YOLO and EfficientNet. This research demonstrates the utilization of both models in mobile applications for the real-time detection of illnesses in rice leaves. The outcomes from the EfficientNet model indicate a value of 1 for all classes, except for brown spot and leaf blast, which both yield a value of 0.99. In contrast, YOLOv8 shows an average accuracy of 0.85–0.92 but excels in inference speed, making it suitable for real-time applications. The best model was then deployed using TensorFlow Lite on the RiceLeaf Detection application, which can provide disease diagnoses directly from rice leaf images. This research demonstrates the potential of lightweight Convolutional Neural Networks (CNNs) to improve early detection of rice diseases, support smart farming practices, and contribute to efforts to improve food security.

Keywords—food security, rice disease detection, Convolutional Neural Network (CNN), EfficientNet, YOLOv8, image processing, TensorFlow Lite, mobile application
 

Cite: Fitri Damayanti, Iwan Santosa, Achmad Imam Sudianto, Imamah, and Sri Herawati, "Sustainable Agriculture: Real-time Detection of Rice Leaf Diseases on Mobile Platforms using Lightweight Deep Learning Models," Journal of Image and Graphics, Vol. 14, No. 5, pp. 879-890, 2026.

Copyright © 2026 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).