Fast method for correcting image misregistration due to organ motion in time-series MRI data

Sandeep N. Gupta, Meiyappan Solaiyappan, Garth M. Beache, Andrew E. Arai, Thomas K.F. Foo

Research output: Contribution to journalArticlepeer-review

39 Scopus citations

Abstract

Time-series MRI data often suffers from image misalignment due to patient movement and respiratory and other physiologic motion during the acquisition process. It is necessary that this misalignment be corrected prior to any automated quantitative analysis. In this article a fast and automated technique for removing in-plane misalignment from time-series MRI data is presented. The method is computationally efficient, robust, and fine-tuned for the clinical setting. The method was implemented and tested on data from 21 human subjects, including myocardial perfusion imaging, renal perfusion imaging, and blood-oxygen level-dependent cardiac T2 imaging. In these applications 10-fold or better reduction in image misalignment is reported. The improvement after registration on representative time-intensity curves is shown. Although the method currently corrects translation motion using image center of mass, the mathematical framework of our approach may be extended to correct rotation and other higher-order displacements.

Original languageEnglish (US)
Pages (from-to)506-514
Number of pages9
JournalMagnetic resonance in medicine
Volume49
Issue number3
DOIs
StatePublished - Mar 1 2003

Keywords

  • BOLD imaging
  • Image registration
  • Motion correction
  • Perfusion imaging

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging

Fingerprint

Dive into the research topics of 'Fast method for correcting image misregistration due to organ motion in time-series MRI data'. Together they form a unique fingerprint.

Cite this