Retinopathy Diagnosis Through Digital Retinal Images

Abstract

Retina is one of the importanttissue in the human eye where the image is focused. There are different layers in retina. Due to excess of hyperglycemia, glucose and other protein deposits accumulate within the retinal layers which causes the vision distorted and blurry. If these changes are not identified in early stages it can cause some serious visual impairments or blindness. There are different techniques to detect retinal pathology but there is one technique which provides accurate and objective visualization of intra-retinal pathology is optical coherence tomography (OCT). There has been some research work done previously on the automated detection of retinal disease from OCT images but as far as we know there is no software based clinical decision system that can predict three different kind of retina diseases. So we are going to introduce world’s first digital clinical decision support system which is going to detect retinal pathology automatically and after that we will diagnose retinal edema(RE), Age related macular edema (AMD) and central serous retinopathy (CSR) based upon the retinal fluid type and morphology using convolutional neural network (CNN)