Automated Screening of Diabetic Retinopathy through Advanced Image Processing Techniques

Abstract Diabetic retinopathy is a retinal vascular disease resulting from fluid leakage in damaged blood vessels. Early diagnosis is essential and can be supported by automated screening systems that rely on reliable image analysis techniques. This study proposes an exudate detection approach employing both coarse and fine segmentation. The coarse stage applies a local variation operator to extract candidate regions with well-defined edges. Validation against clinician-labelled ground truth yielded a sensitivity of 89.7%, specificity of 99.3%, and accuracy of 99.4% on a retinal image dataset, indicating robustness to variations in image quality. Keywords: Diabetic retinopathy; retinal image processing; image analysis; optic disc localization; blood vessel segmentation Citation Shaji C, V. Betcy Thanga Shoba, Jovin R. B & Arockia Rajasekar, “Automated Screening of Diabetic Retinopathy through Advanced Image Processing Techniques”, International Journal of Current Science Research (IJCSR) e-ISSN: 2454-5422: 12(2) 2026: 1-10 License © 2026 The Author(s). Published by Dr. BGR Publications.