A computational study of the focus-of-attention EM-ML algorithm for PET reconstruction

Jens Gregor, Dean A. Huff

Research output: Contribution to journalArticle

6 Citations (Scopus)

Abstract

The expectation-maximization maximum-likelihood (EM-ML) algorithm for image reconstruction in positron emission tomography (PET) essentially solves a large linear system of equations. In this paper, we study computational aspects of a recently developed preprocessing scheme for focusing the attention, and thus the computational resources, on a subset of the equations and unknowns in order to reduce the storage, computation, and communication requirements of the EM-ML algorithm. The approach is completely data-driven and uses no prior anatomic knowledge. The experimental results are obtained from runs on a small network of workstations using simulated phantom data as well as data obtained from a clinical ECAT 921 PET scanner.

Original languageEnglish (US)
Pages (from-to)1481-1497
Number of pages17
JournalParallel Computing
Volume24
Issue number9-10
StatePublished - Sep 1 1998
Externally publishedYes

Fingerprint

Positron Emission Tomography
Positron emission tomography
Expectation Maximization
Maximum likelihood
Maximum Likelihood
Network of Workstations
Linear system of equations
Image Reconstruction
Phantom
Scanner
Image reconstruction
Prior Knowledge
Data-driven
Linear systems
Preprocessing
Unknown
Resources
Subset
Communication
Requirements

All Science Journal Classification (ASJC) codes

  • Software
  • Theoretical Computer Science
  • Hardware and Architecture
  • Computer Networks and Communications
  • Computer Graphics and Computer-Aided Design
  • Artificial Intelligence

Cite this

A computational study of the focus-of-attention EM-ML algorithm for PET reconstruction. / Gregor, Jens; Huff, Dean A.

In: Parallel Computing, Vol. 24, No. 9-10, 01.09.1998, p. 1481-1497.

Research output: Contribution to journalArticle

@article{970af15f67874c17a7addddd0430377f,
title = "A computational study of the focus-of-attention EM-ML algorithm for PET reconstruction",
abstract = "The expectation-maximization maximum-likelihood (EM-ML) algorithm for image reconstruction in positron emission tomography (PET) essentially solves a large linear system of equations. In this paper, we study computational aspects of a recently developed preprocessing scheme for focusing the attention, and thus the computational resources, on a subset of the equations and unknowns in order to reduce the storage, computation, and communication requirements of the EM-ML algorithm. The approach is completely data-driven and uses no prior anatomic knowledge. The experimental results are obtained from runs on a small network of workstations using simulated phantom data as well as data obtained from a clinical ECAT 921 PET scanner.",
author = "Jens Gregor and Huff, {Dean A.}",
year = "1998",
month = "9",
day = "1",
language = "English (US)",
volume = "24",
pages = "1481--1497",
journal = "Parallel Computing",
issn = "0167-8191",
publisher = "Elsevier",
number = "9-10",

}

TY - JOUR

T1 - A computational study of the focus-of-attention EM-ML algorithm for PET reconstruction

AU - Gregor, Jens

AU - Huff, Dean A.

PY - 1998/9/1

Y1 - 1998/9/1

N2 - The expectation-maximization maximum-likelihood (EM-ML) algorithm for image reconstruction in positron emission tomography (PET) essentially solves a large linear system of equations. In this paper, we study computational aspects of a recently developed preprocessing scheme for focusing the attention, and thus the computational resources, on a subset of the equations and unknowns in order to reduce the storage, computation, and communication requirements of the EM-ML algorithm. The approach is completely data-driven and uses no prior anatomic knowledge. The experimental results are obtained from runs on a small network of workstations using simulated phantom data as well as data obtained from a clinical ECAT 921 PET scanner.

AB - The expectation-maximization maximum-likelihood (EM-ML) algorithm for image reconstruction in positron emission tomography (PET) essentially solves a large linear system of equations. In this paper, we study computational aspects of a recently developed preprocessing scheme for focusing the attention, and thus the computational resources, on a subset of the equations and unknowns in order to reduce the storage, computation, and communication requirements of the EM-ML algorithm. The approach is completely data-driven and uses no prior anatomic knowledge. The experimental results are obtained from runs on a small network of workstations using simulated phantom data as well as data obtained from a clinical ECAT 921 PET scanner.

UR - http://www.scopus.com/inward/record.url?scp=0032157682&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0032157682&partnerID=8YFLogxK

M3 - Article

VL - 24

SP - 1481

EP - 1497

JO - Parallel Computing

JF - Parallel Computing

SN - 0167-8191

IS - 9-10

ER -