A New performance evaluation method for face identification - regression analysis of misidentification risk

Wai-han Ho; Paul Watters

doi:10.1109/CVPR.2007.383276

A New performance evaluation method for face identification - regression analysis of misidentification risk

Wai-han Ho, Paul Watters

School of Computing

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

2 Citations (Scopus)

28 Downloads (Pure)

Abstract

The performance of a face identification system varies with its enrollment size. However, most experiments evaluated the performance of algorithms at only one enrollment size with the rank-1 identification rate. The current practice does not demonstrate the usability of algorithms thoroughly. But the problem is, in order to measure identification performance at different sizes, experimenters have to repeat the evaluation with samples of those sizes, which is almost impossible when they are large. Approaches using the Binomial theorem with match and non-match scores have been proposed to estimate performance at different sizes, but as a separate process from the evaluation itself. This paper presents a new way of evaluating identification algorithms that allows the estimating and comparing of performance at different sizes, using the regression analysis of Misidentification Risk.

Original language	English
Title of host publication	Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007)
Editors	Patrick Flynn
Place of Publication	Minneapolis, MN
Publisher	Institute of Electrical and Electronics Engineers (IEEE)
Number of pages	6
ISBN (Print)	1424411807
DOIs	https://doi.org/10.1109/CVPR.2007.383276
Publication status	Published - 2007
Event	IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007) (25th : 2007) - Minneapolis, MN Duration: 19 Jun 2007 → 21 Jun 2007

Conference

Conference	IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007) (25th : 2007)
City	Minneapolis, MN
Period	19/06/07 → 21/06/07

Bibliographical note

Copyright 2007 IEEE. Reprinted from Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007). This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of Macquarie University’s products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected]. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.

Keywords

binomial distribution
estimation theory
face recognition
performance evaluation
regression analysis

Access to Document

10.1109/CVPR.2007.383276

Publisher version (open access)Final published version, 169 KBLicence: Unspecified

Cite this

@inproceedings{b0014f0464734f37aab60fe4d2dc51bc,

title = "A New performance evaluation method for face identification - regression analysis of misidentification risk",

abstract = "The performance of a face identification system varies with its enrollment size. However, most experiments evaluated the performance of algorithms at only one enrollment size with the rank-1 identification rate. The current practice does not demonstrate the usability of algorithms thoroughly. But the problem is, in order to measure identification performance at different sizes, experimenters have to repeat the evaluation with samples of those sizes, which is almost impossible when they are large. Approaches using the Binomial theorem with match and non-match scores have been proposed to estimate performance at different sizes, but as a separate process from the evaluation itself. This paper presents a new way of evaluating identification algorithms that allows the estimating and comparing of performance at different sizes, using the regression analysis of Misidentification Risk.",

keywords = "binomial distribution, estimation theory, face recognition, performance evaluation, regression analysis",

author = "Wai-han Ho and Paul Watters",

note = "Copyright 2007 IEEE. Reprinted from Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007). This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of Macquarie University{\textquoteright}s products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected]. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.; IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007) (25th : 2007) ; Conference date: 19-06-2007 Through 21-06-2007",

year = "2007",

doi = "10.1109/CVPR.2007.383276",

language = "English",

isbn = "1424411807",

editor = "Patrick Flynn",

booktitle = "Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007)",

publisher = "Institute of Electrical and Electronics Engineers (IEEE)",

address = "United States",

}

Ho, W & Watters, P 2007, A New performance evaluation method for face identification - regression analysis of misidentification risk. in P Flynn (ed.), Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007). Institute of Electrical and Electronics Engineers (IEEE), Minneapolis, MN, IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007) (25th : 2007), Minneapolis, MN, 19/06/07. https://doi.org/10.1109/CVPR.2007.383276

A New performance evaluation method for face identification - regression analysis of misidentification risk. / Ho, Wai-han; Watters, Paul.
Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007). ed. / Patrick Flynn. Minneapolis, MN: Institute of Electrical and Electronics Engineers (IEEE), 2007.

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

TY - GEN

T1 - A New performance evaluation method for face identification - regression analysis of misidentification risk

AU - Ho, Wai-han

AU - Watters, Paul

N1 - Copyright 2007 IEEE. Reprinted from Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007). This material is posted here with permission of the IEEE. Such permission of the IEEE does not in any way imply IEEE endorsement of any of Macquarie University’s products or services. Internal or personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution must be obtained from the IEEE by writing to [email protected]. By choosing to view this document, you agree to all provisions of the copyright laws protecting it.

PY - 2007

Y1 - 2007

N2 - The performance of a face identification system varies with its enrollment size. However, most experiments evaluated the performance of algorithms at only one enrollment size with the rank-1 identification rate. The current practice does not demonstrate the usability of algorithms thoroughly. But the problem is, in order to measure identification performance at different sizes, experimenters have to repeat the evaluation with samples of those sizes, which is almost impossible when they are large. Approaches using the Binomial theorem with match and non-match scores have been proposed to estimate performance at different sizes, but as a separate process from the evaluation itself. This paper presents a new way of evaluating identification algorithms that allows the estimating and comparing of performance at different sizes, using the regression analysis of Misidentification Risk.

AB - The performance of a face identification system varies with its enrollment size. However, most experiments evaluated the performance of algorithms at only one enrollment size with the rank-1 identification rate. The current practice does not demonstrate the usability of algorithms thoroughly. But the problem is, in order to measure identification performance at different sizes, experimenters have to repeat the evaluation with samples of those sizes, which is almost impossible when they are large. Approaches using the Binomial theorem with match and non-match scores have been proposed to estimate performance at different sizes, but as a separate process from the evaluation itself. This paper presents a new way of evaluating identification algorithms that allows the estimating and comparing of performance at different sizes, using the regression analysis of Misidentification Risk.

KW - binomial distribution

KW - estimation theory

KW - face recognition

KW - performance evaluation

KW - regression analysis

U2 - 10.1109/CVPR.2007.383276

DO - 10.1109/CVPR.2007.383276

M3 - Conference proceeding contribution

SN - 1424411807

BT - Proceedings of IEEE computer society conference on computer vision and pattern recognition (CVPR 2007)

A2 - Flynn, Patrick

PB - Institute of Electrical and Electronics Engineers (IEEE)

CY - Minneapolis, MN

T2 - IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR 2007) (25th : 2007)

Y2 - 19 June 2007 through 21 June 2007

ER -

A New performance evaluation method for face identification - regression analysis of misidentification risk

Abstract

Conference

Bibliographical note

Keywords

Access to Document

Fingerprint

Cite this