DEEPEYEX: AN ARTIFICIAL INTELLIGENCE-BASED DIAGNOSTIC SUPPORT SYSTEM FOR EYE DISEASES

Tóm tắt

The increasing prevalence of eye diseases in modern society highlights the urgent need for diagnostic solutions that are fast, accurate, and easily accessible. DeepEyeX is proposed as an intelligent healthcare system that supports eye disease diagnosis based on artificial intelligence (AI) and a microservice architecture. The system employs a Convolutional Neural Network (CNN) combined with pre-trained EfficientNet-B3 to analyze retinal images, enabling early detection of diseases such as diabetic retinopathy and glaucoma, while also allowing patients to self-screen for conditions such as conjunctivitis at home. The system architecture includes an admin website for doctors, developed with React Router, and both a web and mobile application for patients, built with Next.js and React Native, supporting appointment booking, diagnosis, and medication purchase. The backend is implemented using Golang (Gin Framework) for business logic processing and Python (FastAPI) for the AI inference service, interconnected through API Gateway Kong within a microservice architecture. Furthermore, the system integrates the Gemini chatbot for automated consultation, VNPAY for online payment, Firebase for authentication and real-time messaging, Stringee for video calls and is containerized and deployed using Docker to ensure flexibility, scalability, and ease of deployment. The PostgreSQL database is securely hosted on AWS RDS, while AWS S3 is used to store retinal images and medical data. Experimental results demonstrate high accuracy in retinal lesion detection, assisting doctors in faster diagnosis and improving patient experience and engagement. In the future, DeepEyeX will be expanded to support more types of eye diseases and integrate medical IoT devices for real-time ocular health monitoring.