Automated Computer Aided Detection of Diabetic Retinopathy using Machine Learning Hybrid Model

Diabetic retinopathy is spreading dangerously worldwide among people with diabetes, leading to reduced vision and completely blindness. In this paper, a technique is proposed to identify diabetic retinopathy using the fundus image obtained from the patient’s retina. The method involves processing the digital image on the fundus image, which help the ophthalmologist in examine DR. Micro-aneurysm, which is considered as first stage of diabetic retinopathy has been diagnosed using Neural Network. The proposed Support Vector Machine has been compared to existing methods - Naive Bayes classifier. Experimental validation is done with MATLAB/SIMULINK software. The preprocess image has been applied as input data for neural network pattern recognition. A significant improvement has been observed in terms of Sensitivity, Specificity and Accuracy compared to the aforementioned existing methods.