Абстрактный

A Novel Approach for Retinal Lesion Detection In Diabetic Retinopathy Images

M. Sridevi Mahe swari , Adarsh Punnolil

In the modern world, diabetic retinopathy (DR) has become one of the most severe complication prevalent among diabetic patients. The success rate of its curability solemnly depends on the early stage diagnosis or else will lead to total blindness. The paper proposes a novel method for the automated identification of exudates pathologies in retinopathy fundus images based on computational intelligence technique. Approach employs a unique sequential execution of morphological operators to extract fundus image features like vessels, red lesions, and white lesions together with texture feature analysis. Finally features selected are passed into the well-known support vector machine (SVM) classifier which classifies the images into normal and abnormal classes. Real time and publicly available database analysis shows really encouraging performance metrics of the proposed method in terms of sensitivity, specificity and accuracy.