AI checker for research papers & theses.

Abstract

Diabetic retinopathy (DR) is a leading cause of preventable blindness among working-age adults, and its burden is rising rapidly in India. Early detection through regular retinal screening can significantly reduce the risk of vision loss; however, access to ophthalmologists remains limited in rural areas. This study explores the potential of deep learning models to support automated DR screening in resource-constrained settings. Using a publicly available dataset of retinal fundus images, we fine-tuned a convolutional neural network to classify images into referable and non-referable DR. The proposed model achieved promising performance, demonstrating its potential as a cost-effective screening tool. The findings highlight the importance of integrating artificial intelligence into primary healthcare to bridge the gap in eye care services.

1. Introduction

Diabetes mellitus has emerged as one of the most pressing public health challenges of the twenty-first century. India, often referred to as the diabetes capital of the world, is home to a large and growing population of people living with diabetes. Among the various complications associated with the disease, diabetic retinopathy stands out as a major cause of visual impairment.

Timely screening plays a crucial role in preventing vision loss, as early-stage DR is often asymptomatic. Nevertheless, conventional screening relies heavily on trained specialists, who are concentrated in urban centres. As a result, a significant proportion of patients in rural areas remain undiagnosed until the disease has progressed to an advanced stage.

In recent years, advances in deep learning have opened new avenues for automated medical image analysis. Several studies have demonstrated that convolutional neural networks can detect DR from fundus photographs with accuracy comparable to that of human experts. Building on this growing body of research, the present study aims to evaluate a lightweight model suitable for deployment on low-cost devices in rural clinics. The remainder of this paper is organised as follows: Section 2 reviews related work, Section 3 describes the methodology, Section 4 presents the results, and Section 5 concludes the paper.