Evaluation of conventional methods for the detection of lesions in diabetic retinopathy images: A research

Abstract

Diabetes Mellitus (DM) is a sporadic ailment which arises with the vagaries in blood glucose levels. Prolonged history of DM will result in retinal vasculature impediments and leads to Diabetic Retinopathy (DR). The patho is characterized by leakage of blood, fat and protein based particles into the macula and instigates the vision problems. The reliability Conventional clinician’s screening methods is dependent on skilled professionals for diagnosis and screening. It costs to a great deal of time with manual labor and hence there is a great need to automate DR detection. Usage of image processing and machine learning approach to sense various retinopathy aberrations gained huge attraction in recent past. This paper reveals various DR detection and classification methods, including tools, implemented techniques and datasets used. It wishes to help researchers by giving brief literature review of merits and demerits of existing methods, so that it will help them to plan future developments.