Convergence study of an accelerated ML-EM algorithm using bigger step size

Do Sik Hwang, Gengsheng L. Zeng

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

In SPECT/PET, the maximum-likelihood expectation-maximization (ML-EM) algorithm is getting more attention as the speed of computers increases. This is because it can incorporate various physical aspects into the reconstruction process leading to a more accurate reconstruction than other analytical methods such as filtered-backprojection algorithms. However, the convergence rate of the ML-EM algorithm is very slow. Several methods have been developed to speed it up, such as the ordered-subset expectation-maximization (OS-EM) algorithm. Even though OS-type algorithms can bring about significant acceleration in the iterative reconstruction, it is generally believed that ML-EM produces better images, in terms of statistical noise in the reconstruction. In this paper, we present an accelerated ML-EM algorithm with bigger step size and show its convergence characteristics in terms of variance noise and log-likelihood values. We also show some advantages of our method over other accelerating methods using additive forms.

Original languageEnglish
Pages (from-to)237-252
Number of pages16
JournalPhysics in medicine and biology
Volume51
Issue number2
DOIs
Publication statusPublished - 2006 Jan 21

Fingerprint

Maximum likelihood
Single-Photon Emission-Computed Tomography
Set theory
set theory
Noise

All Science Journal Classification (ASJC) codes

  • Radiological and Ultrasound Technology
  • Radiology Nuclear Medicine and imaging

Cite this

@article{3c2f6221dcde471680294eea848a5dc0,
title = "Convergence study of an accelerated ML-EM algorithm using bigger step size",
abstract = "In SPECT/PET, the maximum-likelihood expectation-maximization (ML-EM) algorithm is getting more attention as the speed of computers increases. This is because it can incorporate various physical aspects into the reconstruction process leading to a more accurate reconstruction than other analytical methods such as filtered-backprojection algorithms. However, the convergence rate of the ML-EM algorithm is very slow. Several methods have been developed to speed it up, such as the ordered-subset expectation-maximization (OS-EM) algorithm. Even though OS-type algorithms can bring about significant acceleration in the iterative reconstruction, it is generally believed that ML-EM produces better images, in terms of statistical noise in the reconstruction. In this paper, we present an accelerated ML-EM algorithm with bigger step size and show its convergence characteristics in terms of variance noise and log-likelihood values. We also show some advantages of our method over other accelerating methods using additive forms.",
author = "Hwang, {Do Sik} and Zeng, {Gengsheng L.}",
year = "2006",
month = "1",
day = "21",
doi = "10.1088/0031-9155/51/2/004",
language = "English",
volume = "51",
pages = "237--252",
journal = "Physics in Medicine and Biology",
issn = "0031-9155",
publisher = "IOP Publishing Ltd.",
number = "2",

}

Convergence study of an accelerated ML-EM algorithm using bigger step size. / Hwang, Do Sik; Zeng, Gengsheng L.

In: Physics in medicine and biology, Vol. 51, No. 2, 21.01.2006, p. 237-252.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Convergence study of an accelerated ML-EM algorithm using bigger step size

AU - Hwang, Do Sik

AU - Zeng, Gengsheng L.

PY - 2006/1/21

Y1 - 2006/1/21

N2 - In SPECT/PET, the maximum-likelihood expectation-maximization (ML-EM) algorithm is getting more attention as the speed of computers increases. This is because it can incorporate various physical aspects into the reconstruction process leading to a more accurate reconstruction than other analytical methods such as filtered-backprojection algorithms. However, the convergence rate of the ML-EM algorithm is very slow. Several methods have been developed to speed it up, such as the ordered-subset expectation-maximization (OS-EM) algorithm. Even though OS-type algorithms can bring about significant acceleration in the iterative reconstruction, it is generally believed that ML-EM produces better images, in terms of statistical noise in the reconstruction. In this paper, we present an accelerated ML-EM algorithm with bigger step size and show its convergence characteristics in terms of variance noise and log-likelihood values. We also show some advantages of our method over other accelerating methods using additive forms.

AB - In SPECT/PET, the maximum-likelihood expectation-maximization (ML-EM) algorithm is getting more attention as the speed of computers increases. This is because it can incorporate various physical aspects into the reconstruction process leading to a more accurate reconstruction than other analytical methods such as filtered-backprojection algorithms. However, the convergence rate of the ML-EM algorithm is very slow. Several methods have been developed to speed it up, such as the ordered-subset expectation-maximization (OS-EM) algorithm. Even though OS-type algorithms can bring about significant acceleration in the iterative reconstruction, it is generally believed that ML-EM produces better images, in terms of statistical noise in the reconstruction. In this paper, we present an accelerated ML-EM algorithm with bigger step size and show its convergence characteristics in terms of variance noise and log-likelihood values. We also show some advantages of our method over other accelerating methods using additive forms.

UR - http://www.scopus.com/inward/record.url?scp=32544441568&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=32544441568&partnerID=8YFLogxK

U2 - 10.1088/0031-9155/51/2/004

DO - 10.1088/0031-9155/51/2/004

M3 - Article

C2 - 16394336

AN - SCOPUS:32544441568

VL - 51

SP - 237

EP - 252

JO - Physics in Medicine and Biology

JF - Physics in Medicine and Biology

SN - 0031-9155

IS - 2

ER -