Task Submission
Submission of reconstruction algorithms via Docker containers to compete in standardized tasks evaluating accuracy and computational efficiency under high acceleration factors and limited hardware.
AI/CORE Unranked
CMRx4DFlow2026 2026
Ultra-Fast 4D Flow MRI Reconstruction Challenge
CMRx4DFlow2026 is an AI-driven challenge focused on accelerating 4D Flow MRI reconstruction to enable clinically viable, ultra-fast cardiac and vascular imaging. Hosted as part of MICCAI 2026, it aims to overcome long acquisition times by leveraging deep learning to reconstruct high-fidelity hemodynamic data from extreme undersampling (10x–50x) using a multi-center dataset of over 400 cases.
Paper fit
Contribution paths
A strong submission should clearly identify its contribution and evaluate it appropriately.
Challenge Paper
Submission of peer-reviewed manuscripts (8 pages + 2 pages references) describing methods, experiments, and results using the Springer LNCS template, submitted via OpenReview for presentation at the MICCAI workshop.
Research areas in scope
01
Challenge Tasks
Ultra-Fast 4D Flow MRI ReconstructionAI-driven Hemodynamic ReconstructionClinical Adoption of 4D Flow MRIComputational Efficiency in MRI ReconstructionCross-Site and Cross-Organ Generalizability
Policies worth checking twice
- Manuscripts must be anonymized for peer review.
- Papers are limited to 8 pages of main content plus up to 2 pages for references.
- Submissions must describe original work not previously published or simultaneously submitted elsewhere.
- Papers must focus on challenge methods, experimental settings, and validation results.
- At least one author of each accepted paper must register and present at the workshop.
- Task submissions must be delivered via Docker containers on specified hardware (NVIDIA A6000).
Official sources
Compiled from the official call for papers. The organizers’ pages remain authoritative.
Last verified September 9, 2026