Detection of Hard Exudates in Color Fundus Images of the Human Retina

Diabetic Retinopathy is a progressive ocular disease. The disease may advance from mild to severe non-proliferative diabetic retinopathy and it is one of the most significant factors contributing to blindness. Therefore, it is necessary everyone with diabetes should get a comprehensive dilated eye exam at least once in a year. In this study, a state-of-art image processing techniques to automatically detect the presence of hard exudates in the fundus images are presented. After the contrast adaptive histogram equalization as preprocessing stage, contextual clustering algorithms have been applied to segment the exudates. The key features are like the standard deviation, mean, intensity, edge strength and compactness of the segmented regions are extracted and fed as inputs into Echo State Neural Network (ESNN) to discriminate between the normal and pathological image. A total of 50 images have been used to find the exudates out of which 35 images consisting of both normal and abnormal are used to train the ESSN and the remaining 15 images are used to test the neural network. Furthermore, it confirms 93.0% sensitivity and 100% specificity in terms of exudates based classification.