FLARE-Task4-MRI-FM: 10,000+ MRI Scans for Foundation Model Development

10,000+ MRI scans compiled by FLARE-MedFM for foundation model pretraining, as of July 2025. The dataset supports downstream tasks including liver tumor segmentation, cardiac tissue segmentation, and autism diagnosis. It is the official dataset for Task 4 of the MICCAI FLARE25 challenge.

Use Cases

• Pretrain medical imaging foundation models based on the provided 10,000+ MRI scans.
• Fine-tune liver tumor segmentation models based on the ATLAS23_liver_tumor_seg task.
• Develop cardiac tissue and pathology segmentation models based on the EMIDEC_heart_seg_and_class task.
• Train binary classifiers for heart myocardial status (pathological vs normal) based on the EMIDEC_heart_seg_and_class labels.
• Build models for autism diagnosis based on the mentioned autism diagnosis task.

Strengths

• Contains over 10,000 MRI scans for model pretraining.
• Officially supports multiple downstream medical imaging tasks as part of the MICCAI FLARE25 challenge.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count beyond '10,000+' is unknown, which may limit suitability assessment.
• Freshness should be verified as the last metadata update was 2025-07-03.

Provenance

Source

FLARE-MedFM

Freshness

Last updated 2025-07-03 21:55:29.