This dataset builds on the MEDPIX 2.0 dataset to create a Visual Question Answering (VQA) resource for medical imagery. It complements existing datasets like VQA-RAD, which provides domain-expert validated QA pairs from a subset of MEDPIX. While VQA-RAD offers high-quality data, it may not have enough volume for many use cases. This dataset expands on the original captions, topics, and descriptions in MEDPIX by… See the full description on the dataset page: https://huggingface.co/datasets/adishourya/MEDPIX-ShortQA.
