Diabetic Retinopathy Detection using Deep Learning

Diabetic Retinopathy Detection using Deep Learning

S Harshitha Sai, Abdul Sami, D Neha Reddy, G Ganapati Rao

Computer Science and Engineering (Data Science)

Institute of Aeronautical Engineering

Dundigal, Hyderabad

Abstract— Diabetic Retinopathy (DR) is a severe complication of diabetes that can lead to blindness if not detected early. With the growing prevalence of diabetes, timely diagnosis of DR is crucial. Traditional methods rely on manual examination of retinal images by ophthalmologists, which can be time-consuming and error-prone. This project aims to automate DR detection using deep learning, specifically leveraging the ResNet101 architecture. We use the EyePACS dataset, which contains retinal images labeled with five DR severity levels: No DR, Mild DR, Moderate DR, Severe DR, and Proliferative DR.

Our approach uses transfer learning by fine-tuning the pre-trained ResNet101 model to adapt it to the task of DR classification. Data augmentation techniques, such as rotation and flipping, are applied to enhance the model's ability to generalize across different image conditions. The model achieves high accuracy in predicting DR severity, offering reliable predictions with confidence levels. Standard metrics like accuracy, precision, and F1-score are used to evaluate the model’s performance. This automated system can assist healthcare professionals by providing a scalable solution for early DR detection.

 Keywords— Diabetic Retinopathy, Deep Learning, ResNet101, Transfer Learning, EyePACS, Medical Image Classification.