ACE-Net: AutofoCus-Enhanced Convolutional Network for Field Imperfection Estimation with application to high b-value spiral Diffusion MRI

TL;DR

This paper introduces ACE-Net, a deep learning-based method that automatically estimates magnetic field imperfections in high b-value spiral diffusion MRI, enabling artifact-free image reconstruction without external calibration.

Contribution

It presents a novel autofocus-enhanced convolutional network that leverages a compact basis for accurate, automatic field imperfection estimation in rapid MRI imaging.

Findings

Achieved high-quality image reconstruction in high b-value spiral diffusion MRI.

Demonstrated accurate estimation of B0 inhomogeneity and eddy-currents.

Eliminated the need for external calibration in MRI artifact correction.

Abstract

Spatiotemporal magnetic field variations from B0-inhomogeneity and diffusion-encoding-induced eddy-currents can be detrimental to rapid image-encoding schemes such as spiral, EPI and 3D-cones, resulting in undesirable image artifacts. In this work, a data driven approach for automatic estimation of these field imperfections is developed by combining autofocus metrics with deep learning, and by leveraging a compact basis representation of the expected field imperfections. The method was applied to single-shot spiral diffusion MRI at high b-values where accurate estimation of B0 and eddy were obtained, resulting in high quality image reconstruction without need for additional external calibrations.