IMAGE RECOGNITION FOR THE DIAGNOSIS OF DIABETIC RETINOPATHY
Keywords:
Diabetic Retinopathy, Deep Learning, Convolutional Neural Networks (CNN), Medical Image Analysis, ResNet50, VGG19, InceptionV3, Retinal Fundus Images, Streamlit, Flask, Computer-Aided Diagnosis, Healthcare AutomationAbstract
Diabetic Retinopathy (DR) is a major global cause of vision loss that can be prevented, thus, timely and accurate detection is essential for treatment success. This project presents a deep-learning based system for the automated detection and classification of diabetic retinopathy from retinal fundus images. The system deploys a Convolutional Neural Network (CNN), employing transfer learning with models like InceptionV3, ResNet50, or VGG19 to extract defining features. The trained model is integrated into a user-friendly interface built with Python frameworks like Streamlit or Flask, enabling users to upload retinal images and receive instant predictions along with severity grading. The resulting interactive platform is designed to offer ophthalmologists and healthcare staff a scalable, accessible, and a cost-effective screening aid. The proposed system not only demonstrates the potential of deep-learning in medical image analysis but also contributes to reducing the global burden of vision impairment through early and accurate DR detection.
