Detection of Diabetic Retinopathy Using Collaborative Model of CNN with IoMT

Abstract

The cause of blindness that primarily affects middle-aged adults is diabetic retinopathy (DR), due to excessive blood sugar levels. Internet of Medical Things (IoMT) is capable to collect Diabetic Retinopathy-related information remotely using CAD (Computer-aided diagnostic) systems and provide patients with convincing information. Therefore, the primary goal of this study is to identify and categorize the severity of DR fundus images to prevent a diabetic sufferer from going blind. Thus, this paper developed a unique Diabetic Retinopathy Segmentation (DRS) system by fusing the Deep Learning model with optimization techniques. The preprocessing phase of this system is considered to remove noise from the edges. Next, the usable region from the images is extracted using the increasing region segmentation through K-mean clustering. The characteristics of the Area of Interest (AOI) are then extracted and classified into four severity levels using the suggested Hybrid Genetic and Ant Colony Optimization (HGACO) algorithm with the help of a pertained CNN model, Residual Neural Network (RESnet). Additionally, the test of statistical significance evaluates the DRS system’s Segmentation accuracy. The suggested Diabetic Retinopathy System achieves improved categorization outcomes, with sensitivity, accuracy, and specificity numbers.