Original Research

Improved fat–water reconstruction algorithm with graphics hardware acceleration

  1. David H. Johnson PhD1,
  2. Sreenath Narayan MS1,
  3. Chris A. Flask PhD1,2,
  4. David L. Wilson PhD1,2,*

Article first published online: 23 JAN 2010

DOI: 10.1002/jmri.22051

Issue

Journal of Magnetic Resonance Imaging

Journal of Magnetic Resonance Imaging

Volume 31, Issue 2, pages 457–465, February 2010

Additional Information

How to Cite

Johnson, D. H., Narayan, S., Flask, C. A. and Wilson, D. L. (2010), Improved fat–water reconstruction algorithm with graphics hardware acceleration. J. Magn. Reson. Imaging, 31: 457–465. doi: 10.1002/jmri.22051

Author Information

  1. 1

    Department of Biomedical Engineering, Case Western Reserve University, Cleveland, Ohio, USA

  2. 2

    Department of Radiology, Case Western Reserve University and University Hospitals of Cleveland, Cleveland, Ohio, USA

*426 Wickenden Building, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106

Publication History

  1. Issue published online: 23 JAN 2010
  2. Article first published online: 23 JAN 2010
  3. Manuscript Accepted: 2 NOV 2009
  4. Manuscript Received: 22 JUL 2009

Funded by

  • NIH. Grant Numbers: R24CA110943, R01EB004070
  • the National Center for Research Resources, NIH. Grant Number: C06 RR12463-01
  • Interdisciplinary Biomedical Imaging Training Program. Grant Number: T32EB007509

SEARCH

SEARCH BY CITATION

ARTICLE TOOLS

View Full Article (HTML) Get PDF (267K)

Keywords:

  • IDEAL;
  • graphics cards;
  • GPU;
  • fat–water estimation;
  • image reconstruction

Abstract

Purpose:

To develop a fast and robust Iterative Decomposition of water and fat with Echo Asymmetry and Least-squares (IDEAL) reconstruction algorithm using graphics processor unit (GPU) computation.

Materials and Methods:

The fat–water reconstruction was expedited by vectorizing the fat–water parameter estimation, which was implemented on a graphics card to evaluate potential speed increases due to data-parallelization. In addition, we vectorized and compared Brent's method with golden section search for the optimization of the unknown field inhomogeneity parameter (ψ) in the IDEAL equations. The algorithm was made more robust to fat–water ambiguities using a modified planar extrapolation (MPE) of ψ algorithm. As compared to simple planar extrapolation (PE), the use of an averaging filter in MPE made the reconstruction more robust to neighborhoods poorly fit by a two-dimensional plane.

Results:

Fat–water reconstruction time was reduced by up to a factor of 11.6 on a GPU as compared to CPU-only reconstruction. The MPE algorithms incorrectly assigned fewer pixels than PE using careful manual correction as a gold standard (0.7% versus 4.5%; P < 10−4). Brent's method used fewer iterations than golden section search in the vast majority of pixels (6.8 ± 1.5 versus 9.6 ± 1.6 iterations).

Conclusion:

Data sets acquired on a high field scanner can be quickly and robustly reconstructed using our algorithm. A GPU implementation results in significant time savings, which will become increasingly important with the trend toward high resolution mouse and human imaging. J. Magn. Reson. Imaging 2010; 31: 457–465. © 2010 Wiley-Liss, Inc.

View Full Article (HTML) Get PDF (267K)