Automated Identification of Diabetic Retinopathy Using Artificial Neutral Network

An ophthalmic disease that affects the retinal blood vessels called diabetic retinopathy. The diabetic retinopathy results in vision loss. A diabetic retinopathy is not treated in primitive stages may lead to vision loss. The diabetic retinopathy has five different classes. They are normal, mild, moderate, secure, PDR. Generally, highly trained people process the colored fundus image to treat the fatal disease. The manual analysis, and detecting of diabetic retinopathy is complex and even error occurred in results. The manual detection takes long time to diagnose the DR. Using the different computer-based techniques have been used to detect the DR and it shows the retinal blood vessels but it does not differentiate the early stages and unable to process the tedious features. The results from computer vision based gives low accuracy. In this project, Artificial Neural Network (ANN) is used to classify various stages of Diabetic retinopathy. The results obtained from that shows better accuracy and performance.