Abstract

Diabetic retinopathy (DR) is a severe complication of diabetes mellitus and a leading cause of vision impairment and blindness globally. Early detection and accurate classification of DR are critical for preventing severe vision loss. Automated detection systems using supervised learning algorithms have significantly improved the accuracy and efficiency of DR diagnosis and have revolutionized DR detection by leveraging fundus images to classify the severity of the disease. The focus has also expanded to not only detecting the presence of DR but also assessing its severity, thereby enabling tailored treatment strategies. Techniques like transfer learning and ensemble models have shown great promise in refining predictions and addressing the challenges of limited labeled data. Additionally, the exploration of hybrid models that combine traditional machine learning algorithms with deep learning frameworks has further enhanced classification performance, making these systems more robust and efficient. Overall, these advancements underscore a transformative shift in diabetic retinopathy care, moving toward more automated and accurate diagnostic systems that promise improved patient outcomes.In our analysis using the KNN algorithm demonstrated better results, these findings align with current advancements, underscoring the potential of KNN and its variants in enhancing the early detection and classification of diabetic retinopathy. Keywords - Diabetic Retinopathy, KNN, CNN,Deep Learning, Supervised Learning.

Description