Screening Prolifetive Diabetic Retinopathy through color retinal images

Abstract

Diabetic Retinopathy (DR) is a common complication of diabetes that damages the eye's retina. Recognition DR as early as possible is very important to protect patients' vision. We propose a method for screening DR and distinguishing Prolifetive Diabetic Retinopathy (PDR) from Non-Prolifetive Retinopathy (NPDR) automatically through color retinal images. This method evaluates the severity of DR by analyzing the retinal vessel patterns. The retinal vessels are automatically extracted using multiscale matched filters. Then the vessel patterns are analyzed by extracting the vessel net density. The experimental results demonstrate that it is a possible solution to screen DR and distinguish PDR from NPDR by only using color retinal images.