Robust Kernel methods for sparse MR image reconstruction

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

A major challenge in contemporary magnetic resonance imaging (MRI) lies in providing the highest resolution exam possible in the shortest acquisition period. Recently, several authors have proposed the use of L1-norm minimization for the reconstruction of sparse MR images from highly-undersampled k-space data. Despite promising results demonstrating the ability to accurately reconstruct images sampled at rates significantly below the Nyquist criterion, the extensive computational complexity associated with the existing framework limits its clinical practicality. In this work, we propose an alternative recovery framework based on homotopic approximation of the L0-norm and extend the reconstruction problem to a multiscale formulation. In addition to several interesting theoretical properties, practical implementation of this technique effectively resorts to a simple iterative alternation between bilteral filtering and projection of the measured k-space sample set that can be computed in a matter of seconds on a standard PC.

Original languageEnglish (US)
Title of host publicationLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages809-816
Number of pages8
Volume4791 LNCS
EditionPART 1
StatePublished - 2007
Event10th International Conference on Medical Imaging and Computer-Assisted Intervention, MICCAI 2007 - Brisbane, Australia
Duration: Oct 29 2007Nov 2 2007

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
NumberPART 1
Volume4791 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other10th International Conference on Medical Imaging and Computer-Assisted Intervention, MICCAI 2007
CountryAustralia
CityBrisbane
Period10/29/0711/2/07

Fingerprint

Computer-Assisted Image Processing
K-space
Kernel Methods
Robust Methods
Image Reconstruction
Magnetic resonance
Image reconstruction
Computational complexity
Magnetic Resonance Imaging
Imaging techniques
Recovery
L1-norm
Alternation
Computational Complexity
High Resolution
Filtering
Projection
Norm
Formulation
Alternatives

ASJC Scopus subject areas

  • Computer Science(all)
  • Biochemistry, Genetics and Molecular Biology(all)
  • Theoretical Computer Science

Cite this

Trazasko, J. D., Manduca, A., & Borisch, E. (2007). Robust Kernel methods for sparse MR image reconstruction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 1 ed., Vol. 4791 LNCS, pp. 809-816). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4791 LNCS, No. PART 1).

Robust Kernel methods for sparse MR image reconstruction. / Trazasko, Joshua D; Manduca, Armando; Borisch, Eric.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4791 LNCS PART 1. ed. 2007. p. 809-816 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4791 LNCS, No. PART 1).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Trazasko, JD, Manduca, A & Borisch, E 2007, Robust Kernel methods for sparse MR image reconstruction. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 edn, vol. 4791 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 4791 LNCS, pp. 809-816, 10th International Conference on Medical Imaging and Computer-Assisted Intervention, MICCAI 2007, Brisbane, Australia, 10/29/07.
Trazasko JD, Manduca A, Borisch E. Robust Kernel methods for sparse MR image reconstruction. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 ed. Vol. 4791 LNCS. 2007. p. 809-816. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
Trazasko, Joshua D ; Manduca, Armando ; Borisch, Eric. / Robust Kernel methods for sparse MR image reconstruction. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 4791 LNCS PART 1. ed. 2007. pp. 809-816 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).
@inproceedings{add2b5b4faff484b9b05e2ac2af2a4d3,
title = "Robust Kernel methods for sparse MR image reconstruction",
abstract = "A major challenge in contemporary magnetic resonance imaging (MRI) lies in providing the highest resolution exam possible in the shortest acquisition period. Recently, several authors have proposed the use of L1-norm minimization for the reconstruction of sparse MR images from highly-undersampled k-space data. Despite promising results demonstrating the ability to accurately reconstruct images sampled at rates significantly below the Nyquist criterion, the extensive computational complexity associated with the existing framework limits its clinical practicality. In this work, we propose an alternative recovery framework based on homotopic approximation of the L0-norm and extend the reconstruction problem to a multiscale formulation. In addition to several interesting theoretical properties, practical implementation of this technique effectively resorts to a simple iterative alternation between bilteral filtering and projection of the measured k-space sample set that can be computed in a matter of seconds on a standard PC.",
author = "Trazasko, {Joshua D} and Armando Manduca and Eric Borisch",
year = "2007",
language = "English (US)",
isbn = "9783540757566",
volume = "4791 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 1",
pages = "809--816",
booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
edition = "PART 1",

}

TY - GEN

T1 - Robust Kernel methods for sparse MR image reconstruction

AU - Trazasko, Joshua D

AU - Manduca, Armando

AU - Borisch, Eric

PY - 2007

Y1 - 2007

N2 - A major challenge in contemporary magnetic resonance imaging (MRI) lies in providing the highest resolution exam possible in the shortest acquisition period. Recently, several authors have proposed the use of L1-norm minimization for the reconstruction of sparse MR images from highly-undersampled k-space data. Despite promising results demonstrating the ability to accurately reconstruct images sampled at rates significantly below the Nyquist criterion, the extensive computational complexity associated with the existing framework limits its clinical practicality. In this work, we propose an alternative recovery framework based on homotopic approximation of the L0-norm and extend the reconstruction problem to a multiscale formulation. In addition to several interesting theoretical properties, practical implementation of this technique effectively resorts to a simple iterative alternation between bilteral filtering and projection of the measured k-space sample set that can be computed in a matter of seconds on a standard PC.

AB - A major challenge in contemporary magnetic resonance imaging (MRI) lies in providing the highest resolution exam possible in the shortest acquisition period. Recently, several authors have proposed the use of L1-norm minimization for the reconstruction of sparse MR images from highly-undersampled k-space data. Despite promising results demonstrating the ability to accurately reconstruct images sampled at rates significantly below the Nyquist criterion, the extensive computational complexity associated with the existing framework limits its clinical practicality. In this work, we propose an alternative recovery framework based on homotopic approximation of the L0-norm and extend the reconstruction problem to a multiscale formulation. In addition to several interesting theoretical properties, practical implementation of this technique effectively resorts to a simple iterative alternation between bilteral filtering and projection of the measured k-space sample set that can be computed in a matter of seconds on a standard PC.

UR - http://www.scopus.com/inward/record.url?scp=38149040912&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=38149040912&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9783540757566

VL - 4791 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 809

EP - 816

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

ER -