A General Algorithm For Compensation Of Trajectory Errors: Application To Radial Imaging

TL;DR

This paper introduces a novel low-rank matrix optimization algorithm that effectively compensates for trajectory errors in radial MRI imaging, producing artifact-free images without requiring calibration.

Contribution

The paper presents a new low-rank Hankel matrix-based framework for correcting trajectory errors in radial MRI, eliminating the need for prior calibration.

Findings

Successfully removes trajectory-induced artifacts in radial MRI images

Achieves image quality comparable to existing calibration-based methods

Applicable to various radial acquisition schemes including partial Fourier and golden-angle

Abstract

Purpose: To reconstruct artifact-free images from measured k-space data, when the actual k-space trajectory deviates from the nominal trajectory due to gradient imperfections. Methods: Trajectory errors arising from eddy currents and gradient delays introduce phase inconsistencies in several fast scanning MR pulse sequences, resulting in image artifacts. The proposed algorithm provides a novel framework to compensate for this phase distortion. The algorithm relies on the construction of a multi-block Hankel matrix, where each block is constructed from k-space segments with the same phase distortion. In the presence of spatially smooth phase distortions between the segments, the complete block-Hankel matrix is known to be highly low-rank. Since each k-space segment is only acquiring part of the k-space data, the reconstruction of the phase compensated image from their partially parallel measurements is posed as a structured low-rank matrix optimization problem, assuming the coil sensitivities to be known. Results: The proposed formulation is tested on radial acquisitions in several settings including partial Fourier and golden-angle acquisitions. The experiments demonstrate the ability of the algorithm to successfully remove the artifacts arising from the trajectory errors, without the need for trajectory or phase calibration. The quality of the reconstruction was comparable to corrections achieved using the Trajectory Auto-Corrected Image Reconstruction (TrACR) for radial acquisitions. Conclusion: The proposed method provides a general framework for the recovery of artifact-free images from radial trajectories without the need for trajectory calibration.