Domain-conditioned and Temporal-guided Diffusion Modeling for Accelerated Dynamic MRI Reconstruction
Liping Zhang, Iris Yuwen Zhou, Sydney B. Montesi, Li Feng, Fang Liu
Intelligent Imaging Innovation and Translation Lab, Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Massachusetts General Hospital, Harvard Medical School, USA
This repository is the official PyTorch implementation of "Domain-conditioned and Temporal-guided Diffusion Modeling for Accelerated Dynamic MRI Reconstruction" (arxiv, paper). dDiMo enables the diffusion process to characterize spatiotemporal information for time-resolved multi-coil Cartesian and non-Cartesian data and achieves state-of-the-art performance in
- Cartesian-acquired multi-coil cardiac MRI
- Golden-Angle-Radial-acquired multi-coil free-breathing lung MRI
across various undersampling rates.
- January 31, 2025: dDiMo is selected for an ORAL presentation at the scientific sessions of the 2025 ISMRM & ISMRT Annual Meeting & Exhibition in Honolulu Hawai'i 💫
Abstract: This study introduces a domain-conditioned and temporally guided diffusion framework for accelerated dynamic MRI reconstruction, in which the reverse diffusion process is explicitly guided to model spatiotemporal structure in time-resolved data. The framework integrates temporal information from time-resolved dimensions, allowing for the concurrent capture of intraframe spatial features and interframe temporal dynamics in diffusion modeling. Meanwhile, it employs additional spatiotemporal and self-consistent frequency-temporal priors to guide the diffusion process, ensuring precise temporal alignment and enhancing fine image detail recovery. To facilitate a smooth diffusion process, the nonlinear conjugate gradient algorithm is utilized during the reverse diffusion steps. The proposed model was tested on two types of MRI data: Cartesian-acquired multicoil cardiac MRI and golden-angle-radial-acquired multicoil free-breathing lung MRI, across various undersampling rates. It achieved high-quality reconstructions, demonstrating improved temporal alignment and structural recovery compared with other competitive reconstruction methods, both qualitatively and quantitatively. This diffusion framework exhibited robust performance in handling both Cartesian and non-Cartesian acquisitions, effectively reconstructing dynamic datasets in cardiac and lung MRI under different imaging conditions.
See INSTALL.md for the installation of dependencies required to run dDiMo.
To run the dDiMo, you should modify the basic settings in ddimo_cmr.sh for Cartesian-acquired multi-coil cardiac MRI and ddimo_dce.sh for Golden-Angle-Radial-acquired multi-coil free-breathing lung MRI. Detailed default settings can be found in the default_setting folder.
cd cmr_examples/scripts/ddimo_cmr.sh
bash ddimo_cmr.sh
cd dce_examples/scripts/ddimo_dce.sh
bash ddimo_dce.sh
@article{zhang2026domain,
title = {Domain-Conditioned and Temporal-Guided Diffusion Modeling for Accelerated Dynamic MRI Reconstruction},
author = {Zhang, Liping and Yuwen Zhou, Iris and Montesi, Sydney B. and Feng, Li and Liu, Fang},
journal = {NMR in Biomedicine},
volume = {39},
number = {5},
pages = {e70256},
doi = {https://doi.org/10.1002/nbm.70256},
year = {2026},
publisher={Wiley Online Library}
}This project is released under the BSD 2-Clause license. The codes are based on fastMRI, CAMP-Net, ktCLAIR, and GRASP-GROG. Please also follow their licenses. Thanks for their awesome works.