Rice Disease Detection

This work presented an intelligent rice disease detection system that combines deep learning-based classification with explainable AI techniques to improve both accuracy and interpretability. By integrating a Convolutional Neural Network (CNN) with Gradient-weighted Class Activation Mapping (Grad-CAM), the proposed approach addresses the limitation of traditional models that provide predictions without clear explanations. The system is designed as a unified pipeline that performs image preprocessing, disease classification, and visual interpretation within a single framework. Experimental results demonstrate that the model achieves strong classification performance across different rice leaf disease categories, while maintaining consistent reliability under varying conditions. The Grad-CAM module provides meaningful heatmaps that highlight infected regions of the leaf, confirming that the model focuses on relevant features during prediction. This combination of quantitative accuracy and qualitative visualization enhances user confidence and supports better decision-making in agricultural practices. The system architecture enables efficient processing and can be extended to real-time applications, allowing users to upload images and receive immediate results. By automating the detection process, the proposed solution reduces dependency on manual inspection and expert knowledge, making it especially useful for farmers and agricultural stakeholders. Despite these advantages, certain limitations remain. The effectiveness of the model depends on the availability of diverse and high-quality training data. In some cases, variations in lighting, background, or image quality may affect prediction accuracy. Additionally, rare or less-represented diseases may require further data collection and model refinement. Future improvements may include integrating advanced deep learning architectures, incorporating environmental factors such as weather and soil conditions, and deploying the system as a mobile or cloud-based application for wider accessibility. Techniques such as transfer learning, model optimization, and real-time processing can further enhance system performance. In conclusion, the proposed system successfully combines CNN-based classification with Grad-CAM visualization to deliver an accurate, interpretable, and practical solution for rice disease detection. This work contributes to the advancement of smart agriculture by providing a scalable and user-friendly tool for early disease identification and crop management.

[1]  S. Mohanty, D. Hughes, and M. Salathé, “Using deep learning for image-based plant disease detection,” Frontiers in Plant Science, 2016. This work demonstrates the effectiveness of CNN models in classifying plant diseases using large-scale image datasets. [2]   R. Ferentinos, “Deep learning models for plant disease detection and classification,” Computers and Electronics in Agriculture, 2018. The study evaluates multiple CNN architectures for accurate crop disease identification. [3]  S. Sladojevic et al., “Deep neural networks based recognition of plant diseases by leaf image classification,” Computational Intelligence and Neuroscience, 2016. This research highlights automated disease detection using deep neural networks. [4]    H. Hughes and M. Salathé, “An open access repository of plant disease images to enable mobile diagnostics,” arXiv preprint arXiv:1511.08060, 2015. The paper introduces a widely used dataset for plant disease research. [5]       R. Selvaraju et al., “Grad-CAM: Visual explanations from deep networks via gradient-based localization,” Proc. IEEE ICCV, 2017. This work proposes Grad-CAM for visualizing model decision regions. [6]    K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” Proc. IEEE CVPR, 2016. Introduces ResNet architecture for improved deep learning performance. [7]    K. Simonyan and A. Zisserman, “Very deep convolutional networks for large-scale image recognition,” arXiv preprint arXiv:1409.1556, 2014. Presents the VGG architecture for image classification tasks. [8]    A. Krizhevsky, I. Sutskever, and G. Hinton, “ImageNet classification with deep convolutional neural networks,” NeurIPS, 2012. A foundational work in CNN-based image classification. [9]    M. Tan and Q. Le, “EfficientNet: Rethinking model scaling for convolutional neural networks,” ICML, 2019. Proposes an efficient scaling method for CNN models. [10]  J. Too et al., “A comparative study of fine-tuned deep learning models for plant disease identification,” Computers and Electronics in Agriculture, 2019. Compares performance of different CNN models.