DiffCMR: Fast Cardiac MRI Reconstruction with Diffusion Probabilistic Models

Tianqi Xiang,Wenjun Yue,Yiqun Lin,Jiewen Yang,Zhenkun Wang,Xiaomeng Li

2023-12-08

Abstract:Performing magnetic resonance imaging (MRI) reconstruction from under-sampled k-space data can accelerate the procedure to acquire MRI scans and reduce patients' discomfort. The reconstruction problem is usually formulated as a denoising task that removes the noise in under-sampled MRI image slices. Although previous GAN-based methods have achieved good performance in image denoising, they are difficult to train and require careful tuning of hyperparameters. In this paper, we propose a novel MRI denoising framework DiffCMR by leveraging conditional denoising diffusion probabilistic models. Specifically, DiffCMR perceives conditioning signals from the under-sampled MRI image slice and generates its corresponding fully-sampled MRI image slice. During inference, we adopt a multi-round ensembling strategy to stabilize the performance. We validate DiffCMR with cine reconstruction and T1/T2 mapping tasks on MICCAI 2023 Cardiac MRI Reconstruction Challenge (CMRxRecon) dataset. Results show that our method achieves state-of-the-art performance, exceeding previous methods by a significant margin. Code is available at <a class="link-external link-https" href="https://github.com/xmed-lab/DiffCMR" rel="external noopener nofollow">this https URL</a>.

Image and Video Processing,Computer Vision and Pattern Recognition

What problem does this paper attempt to address?

The paper primarily addresses a key issue in the field of Magnetic Resonance Imaging (MRI): how to quickly reconstruct high-quality MRI images from undersampled k-space data. Specifically, the research team proposed a new framework called DiffCMR, which utilizes conditional denoising diffusion probabilistic models (DDPMs) to tackle this problem.

### Main Issues

1. **Accelerating MRI Scans**: MRI is an important non-invasive imaging technique, but its acquisition time is long. This is especially true in cardiac magnetic resonance imaging (CMR), where the uncontrollable beating of the heart requires data acquisition over multiple cardiac cycles. Therefore, shortening MRI scan time can reduce patient discomfort.

2. **Reconstruction of Undersampled MRI Images**: To accelerate the MRI acquisition process, the amount of required data can be reduced by undersampling k-space data. However, this leads to noise and artifacts in the reconstruction process. Thus, effective algorithms are needed to reconstruct high-quality MRI images from these undersampled data.

### Solution

The proposed method, DiffCMR, addresses the above issues through the following key points:

1. **Conditional Denoising Diffusion Probabilistic Models**: Utilizes conditional denoising diffusion probabilistic models to generate fully sampled MRI image slices. The model receives undersampled MRI image slices as conditional signals at each denoising diffusion step.

2. **Multi-round Integration Strategy**: Employs a multi-round integration strategy during the inference phase to stabilize performance.

3. **Experimental Validation**: Extensive experimental validation was conducted on the MICCAI 2023 Cardiac MRI Reconstruction Challenge (CMRxRecon) dataset, showing that the DiffCMR method significantly outperforms existing methods.

In summary, DiffCMR aims to accelerate the MRI image reconstruction process through a novel approach while maintaining or improving image quality, which is of great significance for clinical applications.