Robust Compressed Sensing MRI with Deep Generative Priors

By Ajil Jalal, Marius Arvinte, Giannis Daras, Eric Price, A. Dimakis, Jonathan I. Tamir. Neural Information Processing Systems, 2021

470 citations recorded by the source corpus.

The CSGM framework (Bora-Jalal-Price-Dimakis'17) has shown that deep generative priors can be powerful tools for solving inverse problems. However, to date this framework has been empirically successful only on certain datasets (for example, human faces and MNIST digits), and it is known to perform poorly on out-of-distribution samples. In this paper, we present the first successful application of the CSGM framework on clinical MRI data. We train a generative prior on brain scans from the fastMRI dataset, and show that posterior sampling via Langevin dynamics achieves high quality reconstructions. Furthermore, our experiments and theory show that posterior sampling is robust to changes in the ground-truth distribution and measurement process. Our code and models are available at: \url{https://github.com/utcsilab/csgm-mri-langevin}.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/a9194aa117acabe333aaaf330cf71c66094ced51

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Solving Inverse Problems in Medical Imaging with Score-Based Generative Models (2021, methodology): “The theory of Langevin dynamics guarantees that in the limit of N Ñ 8 and (cid:15) Ñ 0 , x p N q is a sample from p t p x q under some regularity conditions.”
Diffusion models in medical imaging: A comprehensive survey (2022, methodology): “Generative models, specifically diffusion models, can be used as an alternative to these model-based priors since they can provide a more accurate representation of the data distribution [47, 48].”
Deblurring via Stochastic Refinement (2021, background): “These methods perform very well in-distribution, but can be quite fragile to distributional shifts or changes in the corruption process [25, 50].”
Score-based diffusion models for accelerated MRI (2021, background): “made independently in (Jalal et al., 2021), where the authors partially proved that posterior sampling is indeed highly robust to distribution shifts.”
Diffusion Posterior Sampling for General Noisy Inverse Problems (2022, background): “Another line of work (Jalal et al., 2021) solves linear inverse problems where A ( x ) ≜ Ax and utilizes an approximation , which is obtained when n is assumed to be Gaussian noise with variance σ 2 .”
Come-Closer-Diffuse-Faster: Accelerating Conditional Diffusion Models for Inverse Problems through Stochastic Contraction (2021, methodology): “For example, oneiteration of the standard gradient descent [13,24] or projection onto convex sets (POCS) in [9, 11, 27, 35] corresponds to our data consistency step in (14) with (15).”
Denoising Diffusion Restoration Models (2022, background): “This setup is inherently general to all linear inverse problems, a property desired in many real-world applications such as medical imaging [46, 20].”
Unsupervised Medical Image Translation With Adversarial Diffusion Models (2022, methodology): “Given this potential, diffusion-based methods have recently been adopted for unimodal imaging tasks such as image reconstruction [45], [46], [47], [48], [49], unconditional image generation [50], and anomaly detection [51], [52].”