Neural Network Based Method for the Diagnosis of Diabetic Retinopathy

Abstract

Diabetic Retinopathy is a severe and wide-spread eye disease, it is the main cause of blindness for the working age population in western countries. For the diagnosis of Diabetic Retinopathy, digital color fundus images are becoming increasingly important. This fact opens the possibility of applying image processing techniques in order to facilitate and improve diagnosis in different ways. As micro aneurysms are earliest sign of DR, therefore an algorithm able to automatically detect the micro aneurysms in fundus image captured is a necessary preprocessing step for a correct diagnosis. Some methods that address this problem can be found in the literature but they have some drawbacks like accuracy or speed. This system aims to develop and test a new method for detecting the micro aneurysms in retina images. Gray level 2D feature based vessel extraction is done using neural network to do preprocessing. The method is evaluated on DRIVE database and prove to be superior than rule based methods. To identify micro aneurysms in an image morphological opening and image enhancement is performed. A MATLAB implementation of the complete algorithm is developed and tests suggest that the diagnosis in an image can be estimated in shorter time than previous techniques with the same or better accuracy.