Deep Unrolling for Magnetic Resonance Fingerprinting

TL;DR

This paper advances magnetic resonance fingerprinting by developing a deep unrolling method that incorporates physical models and is validated on real non-Cartesian k-space data, improving reliability over prior synthetic-only evaluations.

Contribution

It introduces new encoder choices for the unrolled neural network and evaluates the approach on real-world accelerated MRF scans with non-Cartesian sampling.

Findings

Effective reconstruction on real MRF data with non-Cartesian trajectories

Enhanced model consistency with physical forward models

Improved performance over purely data-driven methods

Abstract

Magnetic Resonance Fingerprinting (MRF) has emerged as a promising quantitative MR imaging approach. Deep learning methods have been proposed for MRF and demonstrated improved performance over classical compressed sensing algorithms. However many of these end-to-end models are physics-free, while consistency of the predictions with respect to the physical forward model is crucial for reliably solving inverse problems. To address this, recently [1] proposed a proximal gradient descent framework that directly incorporates the forward acquisition and Bloch dynamic models within an unrolled learning mechanism. However, [1] only evaluated the unrolled model on synthetic data using Cartesian sampling trajectories. In this paper, as a complementary to [1], we investigate other choices of encoders to build the proximal neural network, and evaluate the deep unrolling algorithm on real accelerated MRF scans with non-Cartesian k-space sampling trajectories.