Multimodal Imaging Dataset for Cerebrovascular Disease

Description

As part of the ARPA-H AIR program, we are collecting prospective and retrospective Computed Tomography Angiography (CTA), Magnetic Resonance Angiography (MRA), and video fluoroscopy recordings. In total, we will collect 2,500 CTA/MRA and 1,500 video fluoroscopy recordings. The 1,500 video fluoroscopy recordings will be paired with the corresponding CTA/MRA data. We will also upload annotations for 150 of the recordings and images and 100 anatomic segmentations. These data will be collected in a multi-institutional effort under the ARPA-H AIR program. The immediate goal of the program is to develop and train autonomous robots for stroke treatment. In the future, this dataset could support autonomous segmentation, labeling, and robotics for a range of cerebrovascular diseases. The data will be uploaded in 4 tranches starting in October 2026 and ending in December 2027.

Update Frequency

Four tranches between October 2026 and December 2027.

License

Creative Commons Attribution 4.0 International License (CC BY 4.0)

Documentation

https://kitware.github.io/TAIR

Managed By

Kitware, Inc.

See all datasets managed by Kitware, Inc..

Contact

rachel.clipp@kitware.com

How to Cite

Multimodal Imaging Dataset for Cerebrovascular Disease was accessed on DATE from https://registry.opendata.aws/arpa-h-air.

Usage Examples

Tutorials

• TAIR Data Tutorial by Kitware, Inc.
  
  

Tools & Applications

• TAIR: Toolkit for Autonomous Interventions and Robots by Kitware