KLASIFIKASI PENYAKIT DIABETIC RETINOPATHY MENGGUNAKAN METODE CONVOLUTIONAL NEURAL NETWORK MODEL RESNET152

Diabetes mellitus, more simply called diabetes, is a serious, long-term condition that occurs when elevated blood glucose levels occur because the body is unable to produce sufficient amounts of the hormone insulin or is unable to use insulin effectively. Diabetic Retinopathy detection can be done manually by an ophthalmologist, and the system can also do it automatically. In a manual system, the analysis and interpretation of retinal fundus images requires an ophthalmologist, but the cost is very high. This study proposes a Deep Learning method, namely the ResNet152 Model to automatically identify Diabetic Retinopathy. This study applies 3 working scenarios that are implemented in the APTOS 2019 dataset. In the second test scenario, ResNet152 with the CLAHE technique is the best compared to other scenarios in obtaining a precision value of 81%, recall of 82%, and F1-Score of 82%. In this study, the ResNet152 model with the CLAHE process and augmentation can reduce overfitting compared to the ResNet152 model without using CLAHE or the proposed augmentation in terms of performance results.