Detection and classification of Diabetic Retinopathy in Retinal Images using ANN

Abstract

Diabetic retinopathy is a Complication of diabetes that causes vision loss if it is not recognized and treated timely. It is characterized by the changes in blood vessels and abnormalities in macular region. To detect these abnormalities manually, ophthalmologists perform pupil dilation which irritates patient eye. To overcome this drawback, image processing technique is used in diabetic retinopathy. And a completely automated system is presented in this paper for the detection and classification of diabetic retinopathy. This paper focuses on Artificial Neural Network (ANN) to detect diabetic retinopathy in retinal fundus images. To develop this proposed system, a detection of micro-aneurysms, exudates and blood vessels is done from retinal fundus images. GLCM is formed using MATLAB function and several features like entropy, homogeneity, area of micro-aneurysms, exudates and blood vessels act as input to ANN. ANN is used to classify retinal images as mild, moderate and higher cases of diabetic retinopathy. In order to classify the DR images, different classes are represented using relevant and significant features.