Robust Compressed Sensing MRI with Deep Generative Priors
Ajil Jalal, Marius Arvinte, Giannis Daras, Eric Price and, Alexandros G. Dimakis, Jonathan I. Tamir

TL;DR
This paper demonstrates that deep generative priors within the CSGM framework can be effectively applied to clinical MRI data, achieving high-quality reconstructions and robustness to distributional shifts.
Contribution
First successful application of CSGM on clinical MRI data using generative priors trained on brain scans, with posterior sampling via Langevin dynamics.
Findings
High-quality MRI reconstructions achieved
Posterior sampling shows robustness to distributional changes
Framework extends to real clinical data
Abstract
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}.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Code & Models
Videos
Taxonomy
TopicsSparse and Compressive Sensing Techniques · Generative Adversarial Networks and Image Synthesis · Medical Image Segmentation Techniques
