Detection of Diabetic Retinopathy by applying Gabor Filter with Convolutional Neural Network

Abstract

Diabetic Retinopathy is an eye disease that damages patient’s retina having diabetes. This disease will affects the blood vessels of retina which lead to vision loss and blindness. By detecting the diabetic retinopathy disease at the early stage and starting the treatment on time will prevent the patient vision loss and blindness. The dataset that has been elected for the research work is squeezed from the Kaggle repository with 2063 images.  In this paper, the classification model for detection of Diabetic Retinopathy using Convolutional Neural Network has been proposed   by presenting retinal fundus images of both eyes retina of same patient instead of using single eye retina of same patient. The accuracy results that are acquired through the classification models with single eye retina of same patient and both eyes retina of same patient are compared and proved that the classification model with both eyes retina of same patient has been improved in accuracy rate.

Keywords: Convolutional Neural Networks, Deep Learning, Diabetes, Diabetic Retinopathy, Gabor Filter.