Spatiotemporal Diffusion Model with Paired Sampling for Accelerated Cardiac Cine MRI
This work addresses image quality issues in cardiac cine MRI for medical imaging applications, representing an incremental improvement over existing methods.
The paper tackled spatial and temporal blurring in deep learning reconstruction for accelerated cardiac cine MRI by developing a spatiotemporal diffusion model with a paired sampling strategy, resulting in sharper tissue boundaries and clearer motion in expert evaluations on clinical data.
Current deep learning reconstruction for accelerated cardiac cine MRI suffers from spatial and temporal blurring. We aim to improve image sharpness and motion delineation for cine MRI under high undersampling rates. A spatiotemporal diffusion enhancement model conditional on an existing deep learning reconstruction along with a novel paired sampling strategy was developed. The diffusion model provided sharper tissue boundaries and clearer motion than the original reconstruction in experts evaluation on clinical data. The innovative paired sampling strategy substantially reduced artificial noises in the generative results.