This project aims to evaluate the performance of different models for the classification of diabetic retinopathy (DR) in fundus images. The reported perfomance metrics are not always consistent in the literature. Our goal is to provide a fair comparison between different models using the same datasets and evaluation protocol.
