BIMCV-R: 8,000 3D CT Scans with Medical Annotations

8,069 3D CT scans and associated medical annotations curated from the BIMCV database for the MICCAI24 paper. The dataset includes approximately 8,000 image-text pairs, representing over 2 million 2D CT slices. It was uploaded by author cyd0806 to Hugging Face in March 2025.

Use Cases

• Train 3D medical image segmentation models based on the annotated CT scans.
• Develop medical image captioning or report generation models based on the image-text pairs.
• Benchmark computer vision algorithms for volumetric data analysis using the 3D CT scans.
• Conduct research on multimodal learning in radiology using the paired scans and annotations.

Strengths

• Large scale of 8,069 3D CT scans.
• Includes associated medical annotations, providing labels for analysis.
• Derived from the established BIMCV database and curated for a peer-reviewed conference (MICCAI24).

Limitations

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

Provenance

Source

BIMCV Database

Collection Method

Curated for the MICCAI24 paper.

Time Range

null

Freshness

Last updated 2025-03-02 06:03:36; freshness should be verified.

Geography

null