Abstract: In modern world, Diabetic Retinopathy (DR) is the leading cause of blindness. DR is a common reason for blindness in the age group of 20-74 years. The main reason for red lesion in working age population is micro aneurysms (MA) and hemorrhages(HE).The first clinical observable lesion indication of diabetic retinopathy are red lesions. So that the detection of red lesions are the main problem for the prescreening system. Here the proposed work gives the automatic detection of red lesion using the new set of shape features, known as dynamic shape features (DSF). DSF parameter gives shape of flooding blood in images and by using this the severity can be analyzed. Another aim of this paper is developing a MATLAB based Graphical User Interface (GUI) which helps to find out the abnormality and severity by the ophthalmologist easily and accurately.
Keywords: Fundus image, MA, HE, Graphical User Interface, Support Vector Machine.