Detection of Diabetic Retinopathy: Comparative Analysis of Different CNN Models

The paper aims to present and highlight the comparison of outcomes obtained from three Convolution Neural Network (CNN) models for Diabetic Retinopathy Detection Using Fundus Images. Diabetes is on the rise due to a significant increase in human intake of processed carbs and sugar, which is in line with the expanding trend. Diabetic Retinopathy is a condition that affects the majority of diabetic individuals and causes them to lose part or all of their eyesight (DR). Here fundus images are used to train different CNN models, which will help us determine and compare several features of retinal fundus images for the autonomous diagnosis of diabetic retinopathy and to differentiate a diabetic eye from a healthy eye. There are several different characteristics that may be retrieved; thus, it is critical to evaluate minute aspects that a person could ignore during a physical inspection in order to get the most efficient features for diabetic retinopathy identification.

Diabetic retinopathy and macular edema are complications of eye that are common among diabetics and cause serious damage to their retina. Diabetic retinopathy is the major cause of blindness in the developed and modern world, according to studies, over 285 million individuals are expected to suffer from diabetic retinopathy, diabetes, or other eye-related disorders that might impair eyesight in the coming years.