Computing Covariances for Mutual Information Coregistration

Paul Bromiley; Daniel Rueckert; Jo Hajnal; Guang-Zhong Yang

Computing Covariances for Mutual Information Coregistration

Paul Bromiley, Daniel Rueckert (Editor), Jo Hajnal (Editor), Guang-Zhong Yang (Editor)

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

Abstract

Mutual Information (MI) has become a popular similarity measure in multi-modal medical image registration since it was first applied to the problem in 1995. This paper describes a method for calculating the covariance matrix for MI coregistration. We derive an expression for the covariance matrix by identifying MI as a biased log-likelihood measure. The validity of this result is then demonstrated through comparison with the results of Monte-Carlo simulations of the coregistration of T1-weighted to T2-weighted synthetic MR scans of the brain. We conclude with some observations on the theoretical basis of MI as a log-likelihood.

Original language	English
Title of host publication	Proc. 8th Medical Image Understanding and Analysis Conference
Editors	Daniel Rueckert, Jo Hajnal, Guang-Zhong Yang
Place of Publication	Imperial College London
Publisher	BMVA Press
Pages	77-80
Number of pages	4
Publication status	Published - 2004
Event	Medical Image Understanding and Analysis 2004 - Imperial College London Duration: 23 Sept 2004 → 24 Sept 2004

Conference

Conference	Medical Image Understanding and Analysis 2004
City	Imperial College London
Period	23/09/04 → 24/09/04

Cite this

@inproceedings{6bb39646543f4579a1f89bace5d30ff6,

title = "Computing Covariances for Mutual Information Coregistration",

abstract = "Mutual Information (MI) has become a popular similarity measure in multi-modal medical image registration since it was first applied to the problem in 1995. This paper describes a method for calculating the covariance matrix for MI coregistration. We derive an expression for the covariance matrix by identifying MI as a biased log-likelihood measure. The validity of this result is then demonstrated through comparison with the results of Monte-Carlo simulations of the coregistration of T1-weighted to T2-weighted synthetic MR scans of the brain. We conclude with some observations on the theoretical basis of MI as a log-likelihood.",

author = "Paul Bromiley and Daniel Rueckert and Jo Hajnal and Guang-Zhong Yang",

year = "2004",

language = "English",

pages = "77--80",

editor = "Daniel Rueckert and Jo Hajnal and Guang-Zhong Yang",

booktitle = "Proc. 8th Medical Image Understanding and Analysis Conference",

publisher = "BMVA Press",

address = "United Kingdom",

note = "Medical Image Understanding and Analysis 2004 ; Conference date: 23-09-2004 Through 24-09-2004",

}

Computing Covariances for Mutual Information Coregistration. / Bromiley, Paul; Rueckert, Daniel (Editor); Hajnal, Jo (Editor) et al.
Proc. 8th Medical Image Understanding and Analysis Conference. ed. / Daniel Rueckert; Jo Hajnal; Guang-Zhong Yang. Imperial College London: BMVA Press, 2004. p. 77-80.

Research output: Chapter in Book/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Computing Covariances for Mutual Information Coregistration

AU - Bromiley, Paul

A2 - Rueckert, Daniel

A2 - Hajnal, Jo

A2 - Yang, Guang-Zhong

A2 - Rueckert, Daniel

A2 - Hajnal, Jo

A2 - Yang, Guang-Zhong

PY - 2004

Y1 - 2004

N2 - Mutual Information (MI) has become a popular similarity measure in multi-modal medical image registration since it was first applied to the problem in 1995. This paper describes a method for calculating the covariance matrix for MI coregistration. We derive an expression for the covariance matrix by identifying MI as a biased log-likelihood measure. The validity of this result is then demonstrated through comparison with the results of Monte-Carlo simulations of the coregistration of T1-weighted to T2-weighted synthetic MR scans of the brain. We conclude with some observations on the theoretical basis of MI as a log-likelihood.

AB - Mutual Information (MI) has become a popular similarity measure in multi-modal medical image registration since it was first applied to the problem in 1995. This paper describes a method for calculating the covariance matrix for MI coregistration. We derive an expression for the covariance matrix by identifying MI as a biased log-likelihood measure. The validity of this result is then demonstrated through comparison with the results of Monte-Carlo simulations of the coregistration of T1-weighted to T2-weighted synthetic MR scans of the brain. We conclude with some observations on the theoretical basis of MI as a log-likelihood.

M3 - Conference contribution

SP - 77

EP - 80

BT - Proc. 8th Medical Image Understanding and Analysis Conference

PB - BMVA Press

CY - Imperial College London

T2 - Medical Image Understanding and Analysis 2004

Y2 - 23 September 2004 through 24 September 2004

ER -

Computing Covariances for Mutual Information Coregistration

Abstract

Conference

Fingerprint

Cite this