A Survey of Emerging Applications of Diffusion Probabilistic Models in MRI

TL;DR

This survey analyzes diffusion probabilistic models (DPMs) for MRI, contrasting discrete-time DDPMs and continuous-time score-based SDEs and explaining their relationships and conditional variants. It systematically reviews DPM-based MRI applications across reconstruction, image generation, translation, segmentation, anomaly detection, and other topics, detailing domain (image vs k-space), tasks, data needs, and code availability. The paper highlights how DPMs address MRI-specific challenges (motion, undersampling, multi-contrast data) with robust posterior sampling, data consistency, and priors, while also addressing limitations such as computational burden and data privacy concerns. Finally, it discusses trends, challenges, and future directions, including accelerated sampling, high-dimensional MRI, priors incorporation, trustworthy AI considerations, and broader organ/task expansion for clinical impact.

Abstract

Diffusion probabilistic models (DPMs) which employ explicit likelihood characterization and a gradual sampling process to synthesize data, have gained increasing research interest. Despite their huge computational burdens due to the large number of steps involved during sampling, DPMs are widely appreciated in various medical imaging tasks for their high-quality and diversity of generation. Magnetic resonance imaging (MRI) is an important medical imaging modality with excellent soft tissue contrast and superb spatial resolution, which possesses unique opportunities for DPMs. Although there is a recent surge of studies exploring DPMs in MRI, a survey paper of DPMs specifically designed for MRI applications is still lacking. This review article aims to help researchers in the MRI community to grasp the advances of DPMs in different applications. We first introduce the theory of two dominant kinds of DPMs, categorized according to whether the diffusion time step is discrete or continuous, and then provide a comprehensive review of emerging DPMs in MRI, including reconstruction, image generation, image translation, segmentation, anomaly detection, and further research topics. Finally, we discuss the general limitations as well as limitations specific to the MRI tasks of DPMs and point out potential areas that are worth further exploration.

Figures (5)

• Figure 1: Development of Diffusion Probabilistic Models in Medical Imaging and Emerging Application in MRI. (a) Pie chart of the medical imaging modality to which DPMs have been applied. (b) Histogram of DPMs in different applications of MRI.
• Figure 2: Diffusion Process and Reverse Process in Diffusion Probabilistic Models (DPMs).
• Figure 3: Two different approaches to realize conditional sampling during the reverse diffusion process of DPMs in MRI reconstruction. It shows how to sample $\mathbf{x}_{t-1}$ from $\mathbf{x}_{t}$ with the acquired data.
• Figure 4: Model design and common tasks of DPMs in MR image generation (a) and image translation (b).
• Figure 5: Organs considered by DPMs in different MRI applications.