Automated Diabetic Retinopathy Detection from Retinal Fundus Images Using Image Processing

Abstract

Diabetic Retinopathy is a disease which occur common for diabetic patients. This complication affects eyes predominantly the retina by damaging the blood vessels of the delicate retinal tissue which is light sensitive. During initial stages diabetic retinopathy may be asymptomatic and gradually develop mild vision problems. There is a need for automatic detection of early stages of Diabetic retinopathy to prevent damages to the eye, from eye vision loss. For early screening and diagnosis of this complication automatic detection is very important so that we can cure it in early stages. Retinal images are captured using a fundus camera which produce fundus image of the retina. Ophthalmologist can analyze these images to predict the nature of the disease its causes and effects. So with these analysis the level of this complication is diagnosed easily. Here we propose a system which will adopt image processing techniques to localize and segment the pathological regions (Blood Vessels, Exudates, etc.) using fundus images. The system will be developed using Matlab, Tenegra, Weka.