mlsanityData to 14 Oct 2026

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

arXiv 2406.16983 ↗First posted 23 Jun 2024eess.IV cs.AI cs.LGcited by 1

On the map: Diffusion models → Medical imaging and segmentation

Abstract (arXiv)

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.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Medical imaging and segmentation, one of 14 directions in Diffusion models. Builds on 12 of its 15 references that are in the map. First posted on arXiv 23 Jun 2024.

Builds on

12 of 15 references.

References · 12

  1. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  2. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  3. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 9,122
  4. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  5. 2018fastMRI: An Open Dataset and Benchmarks for Accelerated MRIcited by 470
  6. 2019ResUNet++: An Advanced Architecture for Medical Image Segmentationcited by 211
  7. 2017Convolutional Recurrent Neural Networks for Dynamic MR Image Reconstructioncited by 105
  8. 2020Advancing machine learning for MR image reconstruction with an open competition: Overview of the 2019 fastMRI challengecited by 59
  9. 2021Measuring Robustness in Deep Learning Based Compressive Sensingcited by 53
  10. 2019Invert to Learn to Invertcited by 18
  11. 2019i-RIM applied to the fastMRI challengecited by 16
  12. 2020Advancing diagnostic performance and clinical usability of neural networks via adversarial training and dual batch normalizationcited by 12

Built on by

Cited by year

Citing papers per year: 2025: 1.

1 paper.

  1. 2025Training-Free Adversarial Robustness in Computational MRIcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Conditional Denoising Diffusion Model-Based Robust MR Image Reconstruction from Highly Undersampled Data
  2. 2023Robust Physics-based Deep MRI Reconstruction Via Diffusion Purification
  3. 2023Guided Reconstruction with Conditioned Diffusion Models for Unsupervised Anomaly Detection in Brain MRIs
  4. 2024Noise Level Adaptive Diffusion Model for Robust Reconstruction of Accelerated MRI
  5. 2023DDM2: Self-Supervised Diffusion MRI Denoising with Generative Diffusion Models
  6. 2022Diffusion Models for Medical Image Analysis: A Comprehensive Survey
  7. 2023Steerable Conditional Diffusion for Out-of-Distribution Adaptation in Medical Image Reconstruction
  8. 2025Elucidating the Design Space of Arbitrary-Noise-Based Diffusion Models