Model observers for assessment of image quality

Harrison H Barrett, Jie Yao, Jannick P. Rolland, Kyle J. Myers

Research output: Contribution to journalArticle

411 Citations (Scopus)

Abstract

Image quality can be defined objectively in terms of the performance of some "observer" (either a human or a mathematical model) for some task of practical interest. If the end user of the image will be a human, model observers are used to predict the task performance of the human as measured by psychophysical studies, and hence to serve as the basis for optimization of image quality. In this paper we consider the task of detection of a weak signal in a noisy image. The mathematical observers considered include the ideal Bayesian, the nonprewhitening matched filter, a model based on linear-discriminant analysis and referred to as the Hotelling observer, and the Hotelling and Bayesian observers modified to account for the spatial-frequency-selective channels in the human visual system. The theory behind these observer models is briefly reviewed, and several psychophysical studies relating to the choice among them are summarized. Only the Hotelling model with channels is mathematically tractable in all cases considered here and capable of accounting for all of these data. This model requires no adjustment of parameters to fit the data and is relatively insensitive to the details of the channel mechanism. We therefore suggest it as a useful model observer for the purpose of assessing and optimizing image quality with respect to simple detection tasks.

Original languageEnglish (US)
Pages (from-to)9758-9765
Number of pages8
JournalProceedings of the National Academy of Sciences of the United States of America
Volume90
Issue number21
StatePublished - Nov 1 1993

Fingerprint

Social Adjustment
Task Performance and Analysis
Discriminant Analysis
Theoretical Models

ASJC Scopus subject areas

  • General
  • Genetics

Cite this

Model observers for assessment of image quality. / Barrett, Harrison H; Yao, Jie; Rolland, Jannick P.; Myers, Kyle J.

In: Proceedings of the National Academy of Sciences of the United States of America, Vol. 90, No. 21, 01.11.1993, p. 9758-9765.

Research output: Contribution to journalArticle

Barrett, Harrison H ; Yao, Jie ; Rolland, Jannick P. ; Myers, Kyle J. / Model observers for assessment of image quality. In: Proceedings of the National Academy of Sciences of the United States of America. 1993 ; Vol. 90, No. 21. pp. 9758-9765.
@article{3b92f122bfe34164b9b15e880b731803,
title = "Model observers for assessment of image quality",
abstract = "Image quality can be defined objectively in terms of the performance of some {"}observer{"} (either a human or a mathematical model) for some task of practical interest. If the end user of the image will be a human, model observers are used to predict the task performance of the human as measured by psychophysical studies, and hence to serve as the basis for optimization of image quality. In this paper we consider the task of detection of a weak signal in a noisy image. The mathematical observers considered include the ideal Bayesian, the nonprewhitening matched filter, a model based on linear-discriminant analysis and referred to as the Hotelling observer, and the Hotelling and Bayesian observers modified to account for the spatial-frequency-selective channels in the human visual system. The theory behind these observer models is briefly reviewed, and several psychophysical studies relating to the choice among them are summarized. Only the Hotelling model with channels is mathematically tractable in all cases considered here and capable of accounting for all of these data. This model requires no adjustment of parameters to fit the data and is relatively insensitive to the details of the channel mechanism. We therefore suggest it as a useful model observer for the purpose of assessing and optimizing image quality with respect to simple detection tasks.",
author = "Barrett, {Harrison H} and Jie Yao and Rolland, {Jannick P.} and Myers, {Kyle J.}",
year = "1993",
month = "11",
day = "1",
language = "English (US)",
volume = "90",
pages = "9758--9765",
journal = "Proceedings of the National Academy of Sciences of the United States of America",
issn = "0027-8424",
number = "21",

}

TY - JOUR

T1 - Model observers for assessment of image quality

AU - Barrett, Harrison H

AU - Yao, Jie

AU - Rolland, Jannick P.

AU - Myers, Kyle J.

PY - 1993/11/1

Y1 - 1993/11/1

N2 - Image quality can be defined objectively in terms of the performance of some "observer" (either a human or a mathematical model) for some task of practical interest. If the end user of the image will be a human, model observers are used to predict the task performance of the human as measured by psychophysical studies, and hence to serve as the basis for optimization of image quality. In this paper we consider the task of detection of a weak signal in a noisy image. The mathematical observers considered include the ideal Bayesian, the nonprewhitening matched filter, a model based on linear-discriminant analysis and referred to as the Hotelling observer, and the Hotelling and Bayesian observers modified to account for the spatial-frequency-selective channels in the human visual system. The theory behind these observer models is briefly reviewed, and several psychophysical studies relating to the choice among them are summarized. Only the Hotelling model with channels is mathematically tractable in all cases considered here and capable of accounting for all of these data. This model requires no adjustment of parameters to fit the data and is relatively insensitive to the details of the channel mechanism. We therefore suggest it as a useful model observer for the purpose of assessing and optimizing image quality with respect to simple detection tasks.

AB - Image quality can be defined objectively in terms of the performance of some "observer" (either a human or a mathematical model) for some task of practical interest. If the end user of the image will be a human, model observers are used to predict the task performance of the human as measured by psychophysical studies, and hence to serve as the basis for optimization of image quality. In this paper we consider the task of detection of a weak signal in a noisy image. The mathematical observers considered include the ideal Bayesian, the nonprewhitening matched filter, a model based on linear-discriminant analysis and referred to as the Hotelling observer, and the Hotelling and Bayesian observers modified to account for the spatial-frequency-selective channels in the human visual system. The theory behind these observer models is briefly reviewed, and several psychophysical studies relating to the choice among them are summarized. Only the Hotelling model with channels is mathematically tractable in all cases considered here and capable of accounting for all of these data. This model requires no adjustment of parameters to fit the data and is relatively insensitive to the details of the channel mechanism. We therefore suggest it as a useful model observer for the purpose of assessing and optimizing image quality with respect to simple detection tasks.

UR - http://www.scopus.com/inward/record.url?scp=0027425561&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0027425561&partnerID=8YFLogxK

M3 - Article

C2 - 8234311

AN - SCOPUS:0027425561

VL - 90

SP - 9758

EP - 9765

JO - Proceedings of the National Academy of Sciences of the United States of America

JF - Proceedings of the National Academy of Sciences of the United States of America

SN - 0027-8424

IS - 21

ER -