Quantitative MRI Scans for Reconstruction and Segmentation

Presenting a subset of the SKM-TEA dataset, containing 2 training scans, 1 validation scan, and 1 test scan. It is intended for tutorials and demos, with a total size of approximately 30GB. The author notes that this mini dataset should not be used for reporting or publishing metrics.

Use Cases

• Demonstrate MRI reconstruction techniques using the provided scan data.
• Build segmentation tutorials based on the training, validation, and test scan splits.
• Create detection demos for quantitative MRI analysis.

Strengths

• Contains a defined split of 2 training, 1 validation, and 1 test scan for structured evaluation.
• Dataset size of approximately 30GB provides substantial data for demonstration purposes.
• Associated with the full SKM-TEA dataset, which is referenced in an arXiv paper (2203.06823).

Limitations

• Very small sample size with only 4 total scans, limiting statistical power.
• Explicitly flagged as unsuitable for reporting or publishing metrics, restricting its use to demos.
• Lacks detailed column descriptions, sample data, and specific file format information.

Provenance

Source

Subset of the SKM-TEA dataset, hosted on Hugging Face by author arjundd.

Collection Method

Subsampled scans from a larger quantitative MRI dataset.

Time Range

null

Freshness

null

Geography

Region is tagged as 'us'.