Automatic Diabetic Assessment for Diabetic Retinopathy Using Support Vector Machines

Diabetic retinopathy become the common disease in most of the countries. Many classifiers are used to classify the extracted features to identify the presence of normal and abnormal retinal images. Microanuerysms, Hemorrhages and exudates are the major symptoms for detecting the proliferative and non-proliferative diabetic retinopathy. In this proposed work, an automated assessment system has been developed for the diabetic retinopathy using the classifier support vector machine which gives the higher accuracy in detecting the micro aneurysms (MAs), hemorrhages (HA) and exudates. In this work, the combination of Support Vector Machine with the Bayesian, probabilistic neural network, clustering, Artificial neural network has been discussed which improves the overall accuracy of the classification.