"Fundus Image Dataset for Five-Stage Diabetic Retinopathy Classification"

"Diabetic retinopathy (DR) is a leading cause of vision impairment worldwide, and accurate automated detection is critical for early intervention. We present a curated dataset of 6,399 retinal fundus images labeled across five DR severity levels: No DR (2,805 images), Mild (740), Moderate (1,898), Proliferative Mild (590), and Severe (366). The dataset captures diverse retinal features and class imbalances typical of real-world clinical scenarios, making it suitable for training and evaluating deep learning models, including convolutional neural networks (CNNs) and vision transformers (ViTs). Extensive annotations and standardized labeling provide a benchmark for research on multi-class DR classification, automated diagnosis, and AI-driven ophthalmic studies. This dataset is intended to facilitate reproducible research and promote the development of robust, clinically relevant DR detection models."