VinDr-CXR-VQA: Chest X-Ray Visual Question Answering with Spatial Grounding

VinDr-CXR-VQA is a large-scale dataset combining 4,394 chest X-ray images with 17,597 natural language question-answer pairs. The dataset, created by faizan711 and last updated in January 2026, is designed for explainable medical AI and includes spatial grounding annotations and clinical reasoning explanations. It features six distinct question types to facilitate research in medical visual question answering.

Use Cases

• Train visual question answering models based on the 17,597 question-answer pairs linked to medical images.
• Develop models for spatial grounding in medical images based on the bounding box annotations mentioned.
• Research explainable AI in radiology based on the provided clinical reasoning explanations.
• Benchmark multimodal AI systems on a medical imaging task based on the combination of images and text.

Strengths

• Contains 4,394 chest X-ray images, providing a substantial visual corpus.
• Includes 17,597 question-answer pairs, offering significant linguistic annotation.
• Features spatial grounding with bounding boxes, enabling localization tasks.
• Provides clinical reasoning explanations, supporting model interpretability.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Derived from the VinDr-CXR dataset.

Collection Method

Likely involves expert annotation of medical images with questions, answers, and bounding boxes.

Freshness

Last updated 2026-01-26 03:33:27; freshness should be verified.