Automated Early Detection of Cardiovascular diseases from Retinal Fundus Images using Predefined Convolution Filters Network

Abstract

The analysis of retinal image is a necessary one to predict different disease in our human body. From the retinal image, cardiovascular diseases can be identified. This can be done by applying feature extraction, segmentation and Classification. Firstly, feature extraction can be done in the retinal by using Gabor filter and then it is segmented by using Particle swarm optimization technique and then it is classified by using Predefined Convolutional Filters Network. The technique used here is applied on online data sets it detects the cardiovascular disease and healthy images with an accuracy of 0.97