Diffusion Models Accelerate Silent Quantitative MRI Mapping
A research paper proposes a diffusion model-based method for quantitative MRI (qMRI) mapping using the MuPa-ZTE sequence, a silent 3D multi-parametric mapping technique with zero echo time. The method combines a denoising diffusion probabilistic model (DDPM) with physics-based data consistency to improve mapping performance and enable acquisition acceleration. It allows high-quality qMRI mapping from a fourfold-accelerated MuPa-ZTE scan, reducing scan time to approximately one minute. The approach leverages the MuPa-ZTE forward signal model as explicit data consistency. The paper is available on arXiv under reference 2512.23726.
Key facts
- MuPa-ZTE is a 3D fast silent multi-parametric mapping sequence with zero echo time.
- MuPa-ZTE uses a 3D phyllotaxis sampling scheme for nearly silent scanning.
- MuPa-ZTE generates quantitative maps of T1, T2, and proton density.
- The proposed method uses a denoising diffusion probabilistic model (DDPM).
- The method incorporates physics-based data consistency.
- It enables fourfold acceleration, reducing scan time to about 1 minute.
- The paper is published on arXiv with ID 2512.23726.
- The method improves patient comfort and motion robustness.
Entities
Institutions
- arXiv