A Segment based Technique for Detecting Exudate from Retinal Fundus Image

Diabetic retinopathy is a disorder that accounting as a root cause of uncorrectable vision loss for diabetic patients. It is due to the damaging of blood vessels that nourish the retina. However, earlier detection and treatment through regular screening, blindness can be avoided. In order to lessen the cost of these screenings, modern image processing techniques are used to voluntarily detect the existence of abnormalities in the retinal images acquired during the screenings. The earliest clinical signs of diabetic retinopathy are microaneurysms, small out-pouching from retinal capillaries, and dot intra-retinal Hemorrhages. Exudates are a major indicator of diabetic retinopathy that can possibly be quantified automatically. Patients having these signs are present in type1 stage of diabetic retinopathy also called NPDR (Non-Proliferative Diabetic Retinopathy). This paper implements a method that identifying the feature of exudates from the image using segment based feature extraction. Here, classification into various stages of NPDR is based on their pixels intensity and frequency from the retinal fundus image. A serious of experiments for extracting the feature is performed with the use of effective image processing techniques. To get these feature value from fundus retinal image various techniques like morphological pre-processing, image boundary tracing, adaptive threshold using Otsu methodology, Optic disk localization are implemented. The SVM classifier uses features extracted by combined 2DPCA instead of explicit image features as the input vector Combined 2DPCA is proposed and then for acquiring higher accuracy of classification we can use virtual SVM. Experimental evaluation on the publicly available dataset DRIVE demonstrates the improved performance of the proposed method for automatic detection of Exudates. These automatically detected exudates are validated by comparing with expert ophthalmologists’ hand-drawn ground-truths. The overall sensitivity of proposed method is 97.1% for the classifier and the specificity is of 98.3%. So, by using this tool Specialist gets support in screening a detection of early changes causing Diabetic Retinopathy and hence timely intervention leading to reduced DR related blindness.