A Conditional Normalizing Flow for Accelerated Multi-Coil MR Imaging

TL;DR

This paper introduces a novel conditional normalizing flow model for accelerated MRI that efficiently samples from the posterior distribution, providing more comprehensive image reconstructions than existing methods.

Contribution

The paper proposes a new CNF model for MRI that infers nullspace components, enabling fast and accurate posterior sampling for accelerated imaging.

Findings

Outperforms recent posterior sampling techniques in MRI reconstruction

Provides faster inference with improved accuracy

Demonstrated on fastMRI brain and knee datasets

Abstract

Accelerated magnetic resonance (MR) imaging attempts to reduce acquisition time by collecting data below the Nyquist rate. As an ill-posed inverse problem, many plausible solutions exist, yet the majority of deep learning approaches generate only a single solution. We instead focus on sampling from the posterior distribution, which provides more comprehensive information for downstream inference tasks. To do this, we design a novel conditional normalizing flow (CNF) that infers the signal component in the measurement operator's nullspace, which is later combined with measured data to form complete images. Using fastMRI brain and knee data, we demonstrate fast inference and accuracy that surpasses recent posterior sampling techniques for MRI. Code is available at https://github.com/jwen307/mri_cnf/