Automated Detection and Classification of Retinal Lesions Using Image Processing Techniques

Abstract

Diabetic Retinopathy (DR), a diabetic eye disease, manifests as damage to the retina due to blood leakage leading to the development of Red Lesions (Microaneurysms and Hemorrhages) and Bright Lesions (Hard Exudates and Cotton Wool Spots). Chronic, uncontrolled diabetes is the primary cause, with delayed treatment potentially resulting in complete blindness. DR is clinically categorized into four stages: No DR, Mild DR, Moderate DR, and Severe DR. Manual detection of DR by ophthalmologists is time-consuming, causing prolonged suffering for patients. To address this, our research leverages advanced technological tools like MATLAB to automate lesion extraction and feature analysis. By quantifying parameters such as lesion number, area, perimeter, and solidity, our model aims to categorize categories accurately. Specifically, we focus on utilizing the area covered by lesions to determine the disease stage, providing valuable insights into disease progression. Our approach accelerates detection and enhances accuracy, potentially improving patient outcomes. By automating lesion extraction and analysis, our system reduces the burden on ophthalmologists, allowing for a more efficient allocation of healthcare resources. This research contributes to the field by offering a robust method for DR assessment, facilitating early intervention and treatment.