Computational analysis and improvement of SIRT

Jens Gregor, Thomas Benson

Research output: Contribution to journalArticle

89 Citations (Scopus)

Abstract

Iterative X-ray computed tomography (CT) algorithms have the potential for producing high-quality images but are computationally very demanding, especially when applied to high-resolution problems. Focusing on simultaneous iterative reconstruction technique (SIRT), we provide an eigenvalue based scheme for automatically determining a near-optimal value of the relaxation parameter. This accelerates the convergence rate of SIRT to the point where only half the number of iterations normally required is needed. We also modify the way SIRT uses preconditioning to solve a weighted least squares problem. The resulting algorithm, which we call PSIRT, is associated with a smaller memory footprint and calls for less data to be communicated in a distributed-memory implementation. Experimental residual norm and timing results are provided based on cone-beam micro-CT mouse data, including for an ordered subsets study.

Original languageEnglish (US)
Article number4492754
Pages (from-to)918-924
Number of pages7
JournalIEEE Transactions on Medical Imaging
Volume27
Issue number7
DOIs
StatePublished - Jul 1 2008
Externally publishedYes

Fingerprint

Tomography
Data storage equipment
Cone-Beam Computed Tomography
X Ray Computed Tomography
Least-Squares Analysis
Image quality
Cones
X rays

All Science Journal Classification (ASJC) codes

  • Software
  • Radiological and Ultrasound Technology
  • Computer Science Applications
  • Electrical and Electronic Engineering

Cite this

Computational analysis and improvement of SIRT. / Gregor, Jens; Benson, Thomas.

In: IEEE Transactions on Medical Imaging, Vol. 27, No. 7, 4492754, 01.07.2008, p. 918-924.

Research output: Contribution to journalArticle

Gregor, Jens ; Benson, Thomas. / Computational analysis and improvement of SIRT. In: IEEE Transactions on Medical Imaging. 2008 ; Vol. 27, No. 7. pp. 918-924.
@article{a1f8a2e2f4c6419c84798abb8c648824,
title = "Computational analysis and improvement of SIRT",
abstract = "Iterative X-ray computed tomography (CT) algorithms have the potential for producing high-quality images but are computationally very demanding, especially when applied to high-resolution problems. Focusing on simultaneous iterative reconstruction technique (SIRT), we provide an eigenvalue based scheme for automatically determining a near-optimal value of the relaxation parameter. This accelerates the convergence rate of SIRT to the point where only half the number of iterations normally required is needed. We also modify the way SIRT uses preconditioning to solve a weighted least squares problem. The resulting algorithm, which we call PSIRT, is associated with a smaller memory footprint and calls for less data to be communicated in a distributed-memory implementation. Experimental residual norm and timing results are provided based on cone-beam micro-CT mouse data, including for an ordered subsets study.",
author = "Jens Gregor and Thomas Benson",
year = "2008",
month = "7",
day = "1",
doi = "10.1109/TMI.2008.923696",
language = "English (US)",
volume = "27",
pages = "918--924",
journal = "IEEE Transactions on Medical Imaging",
issn = "0278-0062",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "7",

}

TY - JOUR

T1 - Computational analysis and improvement of SIRT

AU - Gregor, Jens

AU - Benson, Thomas

PY - 2008/7/1

Y1 - 2008/7/1

N2 - Iterative X-ray computed tomography (CT) algorithms have the potential for producing high-quality images but are computationally very demanding, especially when applied to high-resolution problems. Focusing on simultaneous iterative reconstruction technique (SIRT), we provide an eigenvalue based scheme for automatically determining a near-optimal value of the relaxation parameter. This accelerates the convergence rate of SIRT to the point where only half the number of iterations normally required is needed. We also modify the way SIRT uses preconditioning to solve a weighted least squares problem. The resulting algorithm, which we call PSIRT, is associated with a smaller memory footprint and calls for less data to be communicated in a distributed-memory implementation. Experimental residual norm and timing results are provided based on cone-beam micro-CT mouse data, including for an ordered subsets study.

AB - Iterative X-ray computed tomography (CT) algorithms have the potential for producing high-quality images but are computationally very demanding, especially when applied to high-resolution problems. Focusing on simultaneous iterative reconstruction technique (SIRT), we provide an eigenvalue based scheme for automatically determining a near-optimal value of the relaxation parameter. This accelerates the convergence rate of SIRT to the point where only half the number of iterations normally required is needed. We also modify the way SIRT uses preconditioning to solve a weighted least squares problem. The resulting algorithm, which we call PSIRT, is associated with a smaller memory footprint and calls for less data to be communicated in a distributed-memory implementation. Experimental residual norm and timing results are provided based on cone-beam micro-CT mouse data, including for an ordered subsets study.

UR - http://www.scopus.com/inward/record.url?scp=46349091958&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=46349091958&partnerID=8YFLogxK

U2 - 10.1109/TMI.2008.923696

DO - 10.1109/TMI.2008.923696

M3 - Article

VL - 27

SP - 918

EP - 924

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

SN - 0278-0062

IS - 7

M1 - 4492754

ER -