2003
DOI: 10.1364/josaa.20.001725
|View full text |Cite
|
Sign up to set email alerts
|

Validating the use of channels to estimate the ideal linear observer

Abstract: Image quality can be objectively defined according to how well an observer can perform a task of practical interest given the image. We review a practical model observer for the signal-detection task. The ideal observer for this task is a function of the image probability distributions, which are multidimensional and complicated. This observer is often too difficult to derive or estimate. An alternative to the ideal observer is the ideal linear observer, which can still be unmanageable. Our alternative is the … Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
3
1

Citation Types

1
172
0

Year Published

2012
2012
2021
2021

Publication Types

Select...
6
1

Relationship

0
7

Authors

Journals

citations
Cited by 137 publications
(173 citation statements)
references
References 21 publications
1
172
0
Order By: Relevance
“…One very successful class of model observers are known as Channelized Hotelling observers (CHOs) [2,3], which are an extension of the Hotelling observer that is tractable when the target profile and background statistics are presumed known or can be estimated [4]. CHO models use channels as a set of linear responses, which are then combined to form a decision variable, using Fisher-Hotelling weights.…”
mentioning
confidence: 99%
“…One very successful class of model observers are known as Channelized Hotelling observers (CHOs) [2,3], which are an extension of the Hotelling observer that is tractable when the target profile and background statistics are presumed known or can be estimated [4]. CHO models use channels as a set of linear responses, which are then combined to form a decision variable, using Fisher-Hotelling weights.…”
mentioning
confidence: 99%
“…n I is then estimated in the training phase of the model [7]. Either test statistic (23) or (24) can then be used in order to estimate the area under the ROC (AUC) as a measure of detectability, e.g., using the Wilcoxon-AUC test [16].…”
Section: Estimating the Signal Detectabilitymentioning
confidence: 99%
“…As a general means to estimate the detectability SNR, while alleviating some of the previous problems, numerical observer models have been proposed [1,7]. Some examples are the (non)prewhitening matched filter, the Bayesian ideal observer and the (channelized) Hotelling observer.…”
Section: Introductionmentioning
confidence: 99%
See 2 more Smart Citations