Automatic Detection of Exudate in Diabetic Retinopathy Using K-Clustering Algorithm

Abstract

Diabetic Retinopathy is an eye disease where in which veins may swell and release liquid or new irregular veins develop on retina and piece the light touchy part, this will prompt vision misfortune. It is one of the primary drivers of visual impairment on the planet. Variety in retinal vein thickness, discharge of Exudates which is a protein spillage in the retina, Hemorrhages are a portion of the side effects of Diabetic Retinopathy. Shading fundus pictures will be utilized by ophthalmologists to study eye infections like diabetic retinopathy. Since Optic Disk shows up as a splendid spot in the retinal picture, which takes after exudates, it must be expelled from the picture. Subsequently recognition of Optic Disk is a vital parameter in retinal investigation. On the other hand, in our nation individuals experiencing this disease are all the more in number and therefore oblige more number of ophthalmologists and gigantic time to dissect and analyze the illness. In India, there are insufficient assets, regarding time and accessible master ophthalmologists. In this paper, a programmed and proficient strategy to distinguish Optic Disk and exudates are proposed. The retinal pictures are preprocessed utilizing the method of LAB shading space picture. The preprocessed shading retinal pictures are portioned utilizing Fuzzy C Means grouping method keeping in mind the end goal to distinguish Optic Disk furthermore division is done utilizing Line Operator procedure. Among the over two techniques, best one is recognized. The exudates are removed utilizing K means bunching and finally the grouping is done utilizing SVM. With the characterization accomplished, the Exudates and Non Exudates pictures are separated.
DOI: 10.17762/ijritcc2321-8169.150583