Peningkatan Kontras Citra Retina Mata Menggunakan Ben Graham untuk Indetifikasi Penyakit Diabetes dengan Model ResNet-50

Diabetic retinopathy is a diabetes-related condition that affects the eyes. Early and accurate detection is crucial to improving the chances of recovery. Classifying diabetic retinopathy helps determine the severity level and appropriate treatment. This study aims to enhance image contrast using the Ben Graham method and evaluate its impact on the accuracy of diabetic retinopathy classification using the InceptionV3 model. The dataset used in this research is the “APTOS 2019 Blindness Detection” dataset, consisting of 3,662 images categorized into five classes: No DR, Mild, Moderate, Severe, and Proliferative DR. Two testing scenarios were applied: Scenario 1 with the Ben Graham method and Scenario 2 without it. The test results show an accuracy of 65% for the scenario using Ben Graham and 57% for the scenario without it. These results indicate that applying the Ben Graham method significantly improves the accuracy of diabetic retinopathy classification