Identification and classification of various stages of diabetic retinopathy using convolutional neural network

Abstract

Diabetes causes damage to the retinal blood vessels, resulting in newlineDiabetic Retinopathy (DR). Diabetic Retinopathy (DR) is a serious newlinecomplication of diabetes, and is the prominent root of blindness between newlineindividuals in developed nations. The expected occurrence of diabetes for all newlineage groups worldwide to be 4.4% in 2030. According to projections, there will newlinebe 366 million people with diabetes worldwide by 2030, up from newline171 million in 2000. According to a report from World Health Organization newline(WHO), more than 347 million people worldwide have diabetes, and WHO newlineproposes that diabetes will be the 7th leading cause of death in 2030. The DR newlineis diagnosed by color fundus images, which necessitates the employment of newlinequalified clinicians to detect the presence of lesions. Accurate and timely newlinediagnosis is essential for successful treatment of any disease. newlineComputer Aided Diagnosis (CAD) may prove beneficial for newlinescreening of diseases over a large population and may be time saving as newlinecompared to the physical examination by medical professionals. It will augment newlineand aid the clinical healthcare in the developing countries where there is newlineshortage of trained professional ophthalmologists. The present work proposes newlinea possible solution for DR classification and grading based on fundus image newlineanalysis. Retina consists of several light-sensitive neuron layers, lining the newlineinner surface of the eye, in which many diseases manifest themselves, such as newlinemacular degeneration, glaucoma and diabetic retinopathy. Ophthalmologists newlineand scientists have been seeking the approach to examining the retina for a long newlinetime. newline