Intravoxel B0 inhomogeneity corrected reconstruction using a low-rank encoding operator

Research output: Contribution to journalArticlepeer-review

Abstract

Purpose: To present a general and efficient method for macroscopic intravoxel (Formula presented.) inhomogeneity corrected reconstruction from multi-TE acquisitions. Theory and Methods: A signal encoding model for multi-TE gradient echo (GRE) acquisitions that incorporates 3D intravoxel (Formula presented.) field variations is derived, and a low-rank approximation to the encoding operator is introduced under piecewise linear (Formula presented.) assumption. The low-rank approximation enables very efficient computation and memory usage, and allows the proposed signal model to be integrated into general inverse problem formulations that are compatible with multi-coil and undersampling acquisitions as well as different regularization functions. Results: Experimental multi-echo GRE data were acquired to evaluate the proposed method. Effective reduction of macroscopic intravoxel (Formula presented.) inhomogeneity induced artifacts was demonstrated. Improved (Formula presented.) estimation from the corrected reconstruction over standard Fourier reconstruction has also been obtained. Conclusions: The proposed method can effectively correct the effects of intravoxel (Formula presented.) inhomogeneity, and can be useful for various imaging applications involving GRE-based acquisitions, including fMRI, quantitative (Formula presented.) and susceptibility mapping, and MR spectroscopic imaging.

Original languageEnglish (US)
Pages (from-to)885-894
Number of pages10
JournalMagnetic Resonance in Medicine
Volume84
Issue number2
DOIs
StatePublished - Aug 1 2020

Keywords

  • image reconstruction
  • inhomogeneity correction
  • intravoxel inhomogeneity
  • low-rank models
  • multi-echo acquisitions

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Intravoxel B<sub>0</sub> inhomogeneity corrected reconstruction using a low-rank encoding operator'. Together they form a unique fingerprint.

Cite this