Diabetic Retinopathy Detection Using Deep Learning

The rising prevalence of diabetes has led to an increase in Diabetic Retinopathy (DR), a leading cause of vision loss worldwide. Manual screening of fundus images is time-consuming and prone to human error. This paper proposes a deep learning approach for the automated detection and classification of DR using the EfficientNet-B5 architecture. The system is trained on the APTOS 2019 Blindness Detection dataset, which consists of thousands of high-resolution retinal images. Our model extracts intricate retinal features to classify images into five stages: No DR, Mild, Moderate, Severe, and Proliferative DR. Experimental results demonstrate high diagnostic accuracy, reaching approximately 99.96% confidence in severe cases. The system is deployed via a Flask-based web interface, enabling real-time uploads and automated PDF report generation for clinical support.