Review on: Detection of Diabetic Retinopathy using SVM and MDA

Abstract

Diabetes is the commonest cause of blindness in the working age group these days. Diabetes can affect the sight of a patient and thus it results in causing glaucoma, cataracts and its most severe effect is on blood vessels inside the eye as the blood vessels are damaged and it reaches a condition known as " diabetic retinopathy " which can also be called as eye blindness due to diabetes. Automatic detection of retinal abnormalities is commonly performed for haemorrhages, micro aneurysms, cotton wool spot and hard exudates. However, if more attention is paid to it, there is worse case of retinal abnormality called neovascularisation but much research was not done to detect it. In this case, new blood vessels branch out due to extensive lack of oxygen in the retinal capillaries. So the automated analysis of human eye fundus image is an important task as it can later lead to sectional blindness or thorough blindness. If desired quota of measures is taken and methods are put into consideration then it can be potentially reduced to 50%.In this paper, we present a review for the detection of DR using fundus images and approaches SVM and MDA.