• Suraj Gupta
• Dr. Sneha Soni

Diabetic retinopathy (DR) is a complex issue affecting diabetic patients, leading to retinal damage and potential blindness. This condition disrupts the blood vessels in the retina, causing fluid leakage and severe vision distortion. DR is a prevalent eye disease associated with chronic diabetes and is the leading cause of blindness among working-age adults worldwide, potentially affecting over 93 million individuals. This research introduces an automated classification system capable of analyzing fundus images with varying illumination and fields of view. It employs machine learning models, including Otsu, Random Forest, and Clehe algorithm, to generate severity grades for diabetic retinopathy (DR). The use of machine learning models, such as Random Forest, offers the advantage of high variance and low bias, enabling the classifier to potentially diagnose a broader range of nondiabetic diseases as well. Visualizations of the features learned by the Random Forest classifier and GLMC (Gray Level Matrix Co-occurrence) reveal that the signals used for classification are predominantly located in clearly observable parts of the image. Moderate and severe diabetic retinal images exhibit macroscopic features at a scale that aligns with the architecture's training accuracy and validation accuracy. This research presents a promising approach to automated DR severity classification, offering potential benefits for early diagnosis and intervention in diabetic patients' eye health.

Diabetic Retinopathy, Classification, Image Processing, Deep Learning, Segmentation, Severity Grade

"Efficient and Early Diabetic Retinopathy Detection", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.10, Issue 12, page no.e105-e114, December-2023, Available :http://www.jetir.org/papers/JETIR2312414.pdf

"Efficient and Early Diabetic Retinopathy Detection", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.10, Issue 12, page no. ppe105-e114, December-2023, Available at : http://www.jetir.org/papers/JETIR2312414.pdf