BiomedParseData: Biomedical Image Segmentation and Detection Across Nine Modalities

A processed dataset for training a foundation model for joint segmentation, detection, and recognition of biomedical objects. Each instance includes a 1024x1024 PNG image, a list of textual descriptions for the segmentation target, and a corresponding 1024x1024 binary ground truth mask. The dataset is hosted by Microsoft and was last updated in April 2025.

Use Cases

• Training multimodal vision-language models based on paired biomedical images and textual descriptions.
• Benchmarking segmentation and object detection algorithms based on the provided ground truth masks.
• Developing foundation models for cross-modal biomedical object recognition across nine imaging modalities.

Strengths

• Images and masks are provided at a standardized 1024x1024 resolution.
• Data is processed from multiple public segmentation datasets, suggesting a consolidated source.
• Supports a multimodal foundation model for joint tasks, as described in the associated paper.

Limitations

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

Provenance

Source

Processed from multiple public biomedical segmentation datasets.

Collection Method

Processed and hosted by Microsoft as part of a foundation model research project.

Freshness

Last updated 2025-04-20 21:10:59; freshness should be verified.