This dataset consists of 7,000 fundus images divided into 7 classes: age-related macular degeneration, cataract, diabetic retinopathy, glaucoma, hypertension, myopia, and normal fundus. Each class contains exactly 1,000 images.
All images were annotated using the Qwen2-VL 7B multimodal model for the query: "Describe in detail this image of the fundus with {pathology} pathology. Focus on the associated pathologies that are visible. Use medical terminology and keep the description brief. The… See the full description on the dataset page:
https://huggingface.co/datasets/AlexeyGHT/Fundus_CFP.
