Enhanced Deep Learning model with pipelined preprocessing approach for identifying and classifying Diabetic Retinopathy

Abstract

Diabetic Retinopathy (DR) is a global concern due to its widespread prevalence of diabetes, affecting millions of people worldwide. It causes vision related issues or permanent blindness among the diabetes patients, thereby affecting the quality of their life. The economic burden for treating and diagnosing DR is substantial, affecting both individuals and health care systems. Globalization and Urbanization contributes to higher diabetes rates. Limited access to healthcare in many regions delays diagnosis and treatment. Lack of awareness is another factor that hinders the diabetes management and eye care. Addressing these challenges requires a versatile approach, including awareness, prevention, and improved access to healthcare services. Screening systems are possible potential solutions for the diagnosis of DR as they allow for early detection, slow progression of the disease, cost-effective, and lead to improved population health outcomes. Screening systems are excellent tools to assist Ophthalmologists in the early detection of DR. They improve decision-making accuracy, maintains consistency, minimise clinician workload, and contribute to better patient care. The Objective of this study is to develop efficient screening system strategies that classify the DR severity grading- i.e., No DR, Mild Non-Proliferative Diabetic Retinopathy (NPDR), Moderate NPDR, Severe NPDR, and Proliferative Diabetic Retinopathy (PDR).