Intelligent system for detection and classification of diabetic retinopathy

Diabetic retinopathy is a commonly found medical condition in a diabetic patient due to the exceeding limits of diabetes in blood. A severe case of Diabetic Retinopathy will lead to complete blindness. The main reason Diabetic Retinopathy is dangerous is that it shows no or very few symptoms till it becomes severe and incurable. Early-stage Diabetic Retinopathy detection has a vital role in restoring eye vision and proper treatment. The process of detecting and classifying diabetic retinopathy from retina images is costly and consumes a good amount of time. Also, the detection is done manually by a doctor which makes the detection and classification of the stage prone to errors. An Automated model for the early-stage detection of DR will help the doctors to identify the Diabetic Retinopathy and give proper treatment to those who needed. A few attempts are being done to automate the detection and classification of DR from retina images. This work proposes a machine learning model which detects and classifies diabetic retinopathy from the colored fundus image dataset and classifies them into various stages. The model uses a pre-trained ResNet152 model for feature extraction and classification. An article usually includes an abstract, a concise summary of the work covered at length in the main body of the article. It is used for secondary publications and for information retrieval purposes.