Development of efficient deep learning models for diabetic retinopathy classification using fundus images

Abstract

Deep Learning models have shown remarkable success in a wide newlinevariety of applications like speech recognition, face recognition, computer newlinevision, and much more. Deep Learning models are found to be efficient in newlinelearning complex image representation. In recent days, the success of Deep newlineLearning models in other real-world applications has ignited a way for newlineapplying the same in Medical Imaging. Medical Imaging is one of the most newlinepowerful resources for gaining direct insight into the human body that newlinerepresentations are newlinefurther used in the diagnosis, monitoring, and treatment of medical disorders. newlineThe information provided by each sort of medical imaging technology varies newlinedepending on the disease being analyzed or treated. Medical Imaging can be newlineapplied to capture the details of various internal organs namely lungs, brain, newlinekidneys, retina, and much more. This research focuses on the analysis of newlineretinal disorders which helps ophthalmologist. newlineAmong various retinal diseases, Diabetic Retinopathy (DR), which is a newlineretinal disorder, has caught the attention of various researchers as it has newlinebecome more common in the elderly people or those with a chronic diabetes newlineproblem. The World Health Organization finds that 135 million individuals newlinehave Diabetes worldwide and that the number of individuals with diabetes newlinewill increase to 300 million by the year 2025. Computerized fundus images newlinecaptured using Fundus Camera are utilized for the recognition of Diabetic newlineRetinopathy. Fundus imaging is one of medical technologies used to record newlinethe retinal information from the eyes that may be processed to diagnose newlineretinal disorders. Standard vision screening and color retinal photograph newlineevaluation is needed to identify the condition that creates visual impairment newline