Assessing the accuracy of estimates of the likelihood ratio

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

2 Citations (Scopus)

Abstract

There are many methods to estimate, from ensembles of signal-present and signal-absent images, the area under the receiver operating characteristic curve for an observer in a detection task. For the ideal observer on realistic detection tasks, all of these methods are time consuming due to the difficulty in calculating the ideal-observer test statistic. There are relations, in the form of equations and inequalities, that can be used to check these estimates by comparing them to other quantities that can also be estimated from the ensembles. This is especially useful for evaluating these estimates for any possible bias due to small sample sizes or errors in the calculation of the likelihood ratio. This idea is demonstrated with a simulation of an idealized single photon emission detector array viewing a possible signal in a two-dimensional lumpy activity distribution.

Original languageEnglish (US)
Title of host publicationProceedings of SPIE - The International Society for Optical Engineering
EditorsP. Chakraborty, E.A. Krupinski
Pages135-143
Number of pages9
Volume5034
DOIs
StatePublished - 2003
EventMedical Imaging 2003: Image Perception, Observer Performance, and Technology Assessment - San Diego, CA, United States
Duration: Feb 18 2003Feb 20 2003

Other

OtherMedical Imaging 2003: Image Perception, Observer Performance, and Technology Assessment
CountryUnited States
CitySan Diego, CA
Period2/18/032/20/03

Fingerprint

likelihood ratio
estimates
Photons
Statistics
Detectors
receivers
statistics
detectors
photons
curves
simulation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Condensed Matter Physics

Cite this

Clarkson, E. W., Kupinski, M. A., & Hoppin, J. W. (2003). Assessing the accuracy of estimates of the likelihood ratio. In P. Chakraborty, & E. A. Krupinski (Eds.), Proceedings of SPIE - The International Society for Optical Engineering (Vol. 5034, pp. 135-143) https://doi.org/10.1117/12.480339

Assessing the accuracy of estimates of the likelihood ratio. / Clarkson, Eric W; Kupinski, Matthew A; Hoppin, John W.

Proceedings of SPIE - The International Society for Optical Engineering. ed. / P. Chakraborty; E.A. Krupinski. Vol. 5034 2003. p. 135-143.

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

Clarkson, EW, Kupinski, MA & Hoppin, JW 2003, Assessing the accuracy of estimates of the likelihood ratio. in P Chakraborty & EA Krupinski (eds), Proceedings of SPIE - The International Society for Optical Engineering. vol. 5034, pp. 135-143, Medical Imaging 2003: Image Perception, Observer Performance, and Technology Assessment, San Diego, CA, United States, 2/18/03. https://doi.org/10.1117/12.480339
Clarkson EW, Kupinski MA, Hoppin JW. Assessing the accuracy of estimates of the likelihood ratio. In Chakraborty P, Krupinski EA, editors, Proceedings of SPIE - The International Society for Optical Engineering. Vol. 5034. 2003. p. 135-143 https://doi.org/10.1117/12.480339
Clarkson, Eric W ; Kupinski, Matthew A ; Hoppin, John W. / Assessing the accuracy of estimates of the likelihood ratio. Proceedings of SPIE - The International Society for Optical Engineering. editor / P. Chakraborty ; E.A. Krupinski. Vol. 5034 2003. pp. 135-143
@inproceedings{2689fc909b794d6db1c39c1b7db8d128,
title = "Assessing the accuracy of estimates of the likelihood ratio",
abstract = "There are many methods to estimate, from ensembles of signal-present and signal-absent images, the area under the receiver operating characteristic curve for an observer in a detection task. For the ideal observer on realistic detection tasks, all of these methods are time consuming due to the difficulty in calculating the ideal-observer test statistic. There are relations, in the form of equations and inequalities, that can be used to check these estimates by comparing them to other quantities that can also be estimated from the ensembles. This is especially useful for evaluating these estimates for any possible bias due to small sample sizes or errors in the calculation of the likelihood ratio. This idea is demonstrated with a simulation of an idealized single photon emission detector array viewing a possible signal in a two-dimensional lumpy activity distribution.",
author = "Clarkson, {Eric W} and Kupinski, {Matthew A} and Hoppin, {John W.}",
year = "2003",
doi = "10.1117/12.480339",
language = "English (US)",
volume = "5034",
pages = "135--143",
editor = "P. Chakraborty and E.A. Krupinski",
booktitle = "Proceedings of SPIE - The International Society for Optical Engineering",

}

TY - GEN

T1 - Assessing the accuracy of estimates of the likelihood ratio

AU - Clarkson, Eric W

AU - Kupinski, Matthew A

AU - Hoppin, John W.

PY - 2003

Y1 - 2003

N2 - There are many methods to estimate, from ensembles of signal-present and signal-absent images, the area under the receiver operating characteristic curve for an observer in a detection task. For the ideal observer on realistic detection tasks, all of these methods are time consuming due to the difficulty in calculating the ideal-observer test statistic. There are relations, in the form of equations and inequalities, that can be used to check these estimates by comparing them to other quantities that can also be estimated from the ensembles. This is especially useful for evaluating these estimates for any possible bias due to small sample sizes or errors in the calculation of the likelihood ratio. This idea is demonstrated with a simulation of an idealized single photon emission detector array viewing a possible signal in a two-dimensional lumpy activity distribution.

AB - There are many methods to estimate, from ensembles of signal-present and signal-absent images, the area under the receiver operating characteristic curve for an observer in a detection task. For the ideal observer on realistic detection tasks, all of these methods are time consuming due to the difficulty in calculating the ideal-observer test statistic. There are relations, in the form of equations and inequalities, that can be used to check these estimates by comparing them to other quantities that can also be estimated from the ensembles. This is especially useful for evaluating these estimates for any possible bias due to small sample sizes or errors in the calculation of the likelihood ratio. This idea is demonstrated with a simulation of an idealized single photon emission detector array viewing a possible signal in a two-dimensional lumpy activity distribution.

UR - http://www.scopus.com/inward/record.url?scp=0041858097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0041858097&partnerID=8YFLogxK

U2 - 10.1117/12.480339

DO - 10.1117/12.480339

M3 - Conference contribution

AN - SCOPUS:0041858097

VL - 5034

SP - 135

EP - 143

BT - Proceedings of SPIE - The International Society for Optical Engineering

A2 - Chakraborty, P.

A2 - Krupinski, E.A.

ER -