Smart Plant Disease Diagnosis via MERN Stack Interface and PyTorch Deep Learning Models
Context: Early identification of plant diseases plays a crucial role in enhancing crop productivity and promoting sustainable agricultural practices. Advances in artificial intelligence and web-based technologies have paved the way for smart systems capable of automatically diagnosing diseases in crops like rice and sugarcane. Objective: This research focuses on developing a smart plant disease diagnosis system that integrates deep learning techniques with a MERN (MongoDB, Express.js, React.js, Node.js) stack to provide accurate, real-time classification of rice and sugarcane leaves diseases through a user-friendly web interface. Method: The proposed framework employs a Convolutional Neural Network (CNN) built with PyTorch and trained using a carefully curated dataset of diseased rice and sugarcane leaf images. The developed model was incorporated into a web application built using the MERN stack to enable seamless frontend-backend communication and real-time disease prediction. The model’s effectiveness was assessed using evaluation metrics such as precision, recall, F1-score, and confusion matrix analysis. Results: The CNN model achieved high classification performance, with an average class accuracy of 95.92%, overall classification accuracy of 91.83%, average precision of 91.85%, average recall of 92.05%, and average F1-score of 91.86%. Confusion matrix analysis further validated the model’s efficiency in accurately recognizing rice and sugarcane leaves diseases. The integrated web platform demonstrated efficient and user-friendly real-time disease prediction capabilities. Conclusions: The developed AI-based plant disease detection system highlights the effectiveness of integrating deep learning techniques with modern web technologies to support scalable agricultural solutions. The system provides a practical solution for farmers and agronomists seeking early and accurate crop disease detection. Future enhancements may include multilingual support, mobile application integration, and agronomic advisory modules to further advance precision agriculture initiatives.