21,729 3D medical images and 143,518 corresponding masks comprise this dataset, which was created to facilitate the development of general-purpose foundation models for 3D medical image segmentation. It was gathered from a combination of 70 public datasets and 8,128 privately licensed annotated cases from 24 hospitals. The dataset was authored by blueyo0 and last updated on 2025-05-25.
Use Cases
- Training general-purpose foundation models for 3D medical image segmentation based on the dataset's stated purpose.
- Benchmarking segmentation algorithms on a large-scale, multi-anatomical dataset.
- Developing multi-modal segmentation models based on the dataset's multi-modal nature.
- Conducting research on domain adaptation or generalization across different anatomical regions and imaging sources.
Strengths
- Large scale with 21,729 3D images and 143,518 masks.
- Multi-source composition from 70 public datasets and 8,128 private cases.
- Multi-anatomical and multi-modal scope, as described.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Data may reflect geographic or institutional bias inherent to its source from 24 hospitals.
Provenance
- Source
- Combination of 70 public datasets and 8,128 privately licensed cases from 24 hospitals.
- Collection Method
- Gathered and aggregated from multiple sources.
- Time Range
- null
- Freshness
- Last updated 2025-05-25 04:33:53; freshness should be verified.
- Geography
- null