CDiffMR: Can We Replace the Gaussian Noise with K-Space Undersampling for Fast MRI?

TL;DR

This paper introduces CDiffMR, a novel MRI reconstruction method replacing Gaussian noise with k-space undersampling in diffusion models, achieving fast and high-quality MRI reconstructions with reusability across different undersampling rates.

Contribution

Proposes a Cold Diffusion-based MRI reconstruction method that uses k-space undersampling instead of Gaussian noise, enabling faster inference and model reusability across undersampling rates.

Findings

Achieves comparable or superior reconstruction quality to state-of-the-art methods.

Reaches inference times of only 1.6 to 3.4% of diffusion model counterparts.

Demonstrates effective reusability of the pre-trained model for different undersampling rates.

Abstract

Deep learning has shown the capability to substantially accelerate MRI reconstruction while acquiring fewer measurements. Recently, diffusion models have gained burgeoning interests as a novel group of deep learning-based generative methods. These methods seek to sample data points that belong to a target distribution from a Gaussian distribution, which has been successfully extended to MRI reconstruction. In this work, we proposed a Cold Diffusion-based MRI reconstruction method called CDiffMR. Different from conventional diffusion models, the degradation operation of our CDiffMR is based on \textit{k}-space undersampling instead of adding Gaussian noise, and the restoration network is trained to harness a de-aliaseing function. We also design starting point and data consistency conditioning strategies to guide and accelerate the reverse process. More intriguingly, the pre-trained CDiffMR model can be reused for reconstruction tasks with different undersampling rates. We demonstrated, through extensive numerical and visual experiments, that the proposed CDiffMR can achieve comparable or even superior reconstruction results than state-of-the-art models. Compared to the diffusion model-based counterpart, CDiffMR reaches readily competing results using only $1.6 \sim 3.4\%$ for inference time. The code is publicly available at https://github.com/ayanglab/CDiffMR.