Artificial Intelligence and Deep Learning-Based System Design for Diabetic Retinopathy Classification

Abstract

One of the biggest causes of avoidable blindness throughout the world is diabetic retinopathy (DR). There is a significant unmet need to test all diabetes patients for DR, and many instances of DR go undetected and untreated. In order to automate DR screening, this research aimed to create reliable diagnostic technologies. In order to reduce the pace of vision loss, it is important to refer eyes suspected of having DR to an ophthalmologist for further assessment and treatment. The primary goal of this research is to improve the classification accuracy for Diabetic Retinopathy (DR). In this script, we present a new neural network model for DR forecasting. The suggested model's accuracy in identifying DR phases was measured against that of regular and ensemble-based models. Various benchmark datasets, including MESSIDOR, IDRID, and APTOS, are used in the studies. The suggested DRPNN algorithm outperformed the competition in experiments assessed using industry-standard criteria.