CHAOS: Combined CT and MRI Abdominal Organ Segmentation Challenge Data

CHAOS provides 40 abdominal CT scans and 40 MRI scans from different patients for an organ segmentation challenge. The dataset, created by researchers from Dokuz Eylül University and published in 2019, includes 20 training and 20 testing cases for each modality. Ground truth segmentation masks are provided for the training set, with annotations for the liver in CT scans and for the liver, kidneys, and spleen in MRI scans.

Use Cases

• Train multi-modal segmentation models based on the provided paired CT and MRI training data.
• Benchmark segmentation algorithm performance on healthy abdominal organs using the challenge's test set.
• Develop domain adaptation techniques for medical imaging based on the different annotation schemes between CT and MRI data.
• Pre-train models for abdominal organ detection using the provided ground truth masks for liver, kidneys, and spleen.

Strengths

• Provides 20 training and 20 testing cases for both CT and MRI modalities, enabling multi-modal analysis.
• Includes ground truth segmentation masks for the training data, with annotations for up to four organs.
• MRI data contains both T1-Dual and T2 SPIR sequences, offering multiple contrast images per case.

Limitations

• Ground truth masks for the test set are not publicly available, limiting independent validation.
• The description does not specify the total number of images, file formats beyond DICOM, or dataset size.
• Data may reflect the specific acquisition protocols and patient population of the challenge source.

Provenance

Source

Zenodo, via the CHAOS Grand Challenge.

Collection Method

Collected for the Combined (CT-MR) Healthy Abdominal Organ Segmentation (CHAOS) Challenge.

Time Range

null

Freshness

Last update date is unknown; freshness unverified.

Geography

null