Robust Physics-based Deep MRI Reconstruction Via Diffusion Purification

TL;DR

This paper introduces a diffusion model-based method to enhance the robustness of deep learning MRI reconstruction against measurement perturbations and setting variations, avoiding complex adversarial training.

Contribution

It proposes a novel diffusion purification strategy that improves robustness of MRI reconstruction models without requiring minimax optimization, only fine-tuning on purified data.

Findings

Significantly reduces instabilities compared to adversarial training and randomized smoothing.

Effective in handling worst-case measurement perturbations.

Improves resilience to variations in acceleration factors and sampling locations.

Abstract

Deep learning (DL) techniques have been extensively employed in magnetic resonance imaging (MRI) reconstruction, delivering notable performance enhancements over traditional non-DL methods. Nonetheless, recent studies have identified vulnerabilities in these models during testing, namely, their susceptibility to (\textit{i}) worst-case measurement perturbations and to (\textit{ii}) variations in training/testing settings like acceleration factors and k-space sampling locations. This paper addresses the robustness challenges by leveraging diffusion models. In particular, we present a robustification strategy that improves the resilience of DL-based MRI reconstruction methods by utilizing pretrained diffusion models as noise purifiers. In contrast to conventional robustification methods for DL-based MRI reconstruction, such as adversarial training (AT), our proposed approach eliminates the need to tackle a minimax optimization problem. It only necessitates fine-tuning on purified examples. Our experimental results highlight the efficacy of our approach in mitigating the aforementioned instabilities when compared to leading robustification approaches for deep MRI reconstruction, including AT and randomized smoothing.