On Instabilities of Unsupervised Denoising Diffusion Models in Magnetic Resonance Imaging Reconstruction

TL;DR

This paper reveals that unsupervised diffusion models for MRI reconstruction are vulnerable to small perturbations, which can cause misleading tissue structures and noise artifacts, raising concerns about their clinical reliability.

Contribution

It demonstrates the transferability of worst-case perturbations in diffusion models and highlights their robustness issues in MRI reconstruction.

Findings

Tiny perturbations can cause fake tissue structures.

Larger perturbations produce noise-like artifacts.

Diffusion models are more vulnerable than supervised models.

Abstract

Denoising diffusion models offer a promising approach to accelerating magnetic resonance imaging (MRI) and producing diagnostic-level images in an unsupervised manner. However, our study demonstrates that even tiny worst-case potential perturbations transferred from a surrogate model can cause these models to generate fake tissue structures that may mislead clinicians. The transferability of such worst-case perturbations indicates that the robustness of image reconstruction may be compromised due to MR system imperfections or other sources of noise. Moreover, at larger perturbation strengths, diffusion models exhibit Gaussian noise-like artifacts that are distinct from those observed in supervised models and are more challenging to detect. Our results highlight the vulnerability of current state-of-the-art diffusion-based reconstruction models to possible worst-case perturbations and underscore the need for further research to improve their robustness and reliability in clinical settings.