On Instabilities of Unsupervised Denoising Diffusion Models in Magnetic Resonance Imaging Reconstruction
This is an incremental finding that highlights a critical vulnerability in state-of-the-art diffusion models for MRI reconstruction, posing risks for clinical diagnosis.
The study tackled the problem of unsupervised denoising diffusion models in MRI reconstruction being vulnerable to worst-case perturbations, resulting in the generation of fake tissue structures that could mislead clinicians and Gaussian noise-like artifacts at larger perturbation strengths.
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.