Diabetic Retinopathy Detection & Classification using Efficient Net Model

Diabetic retinopathy is an eye disease that progressively degrades a person’s vision. It affects 40-45% of people with diabetes and it is one of society’s leading causes of blindness. Diabetic retinopathy has 4 different phases and with each stage, the eyesight of a person degrades. If diabetic retinopathy is detected in its early phases its effects and progression can be slowed down and save the person from permanent blindness. We aim to utilize ML to detect the disease and classify diabetic retinopathy into its 4 classes based on severity, helping in early detection. This paper gives a glimpse into diabetic retinopathy and proposes a methodology to develop a machine-learning model along with deploying the developed model on an AWS based web-app.