Clinician-generated question-answer pairs paired with radiology images across open-ended and binary 'yes/no' categories. The dataset utilizes medical imagery sourced from the MedPix open-access database to support the development of Medical Visual Question Answering (VQA) systems.
Use Cases
- Train a medical VQA model to generate text responses based on radiology images and question strings
- Evaluate binary classification performance using the 'yes/no' question subset
- Fine-tune multimodal vision-language models using clinician-validated answer labels
Strengths
- Includes both open-ended and binary 'yes/no' question formats
- Features radiology images sourced from the MedPix open-access database
- Contains question-answer pairs manually generated by a team of clinicians
- Designed specifically for training and testing Medical Visual Question Answering (VQA) systems