A Bayesian model for predicting face recognition performance using image quality

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

7 Citations (Scopus)
78 Downloads (Pure)

Abstract

Quality of a pair of facial images is a strong indicator of the uncertainty in decision about identity based on that image pair. In this paper, we describe a Bayesian approach to model the relation between image quality (like pose, illumination, noise, sharpness, etc) and corresponding face recognition performance. Experiment results based on the MultiPIE data set show that our model can accurately aggregate verification samples into groups for which the verification performance varies fairly consistently. Our model does not require similarity scores and can predict face recognition performance using only image quality information. Such a model has many applications. As an illustrative application, we show improved verification performance when the decision threshold automatically adapts according to the quality of facial images.
Original languageUndefined
Title of host publication2014 IEEE International Joint Conference on Biometrics (IJCB)
Place of PublicationUSA
PublisherIEEE
Pages1-8
Number of pages8
ISBN (Print)978-1-4799-3584-0
DOIs
Publication statusPublished - 29 Sep 2014
Event2014 IEEE International Joint Conference on Biometrics, IJCB 2014 - Clearwater, United States
Duration: 29 Sep 20142 Oct 2014

Publication series

Name
PublisherIEEE

Conference

Conference2014 IEEE International Joint Conference on Biometrics, IJCB 2014
Abbreviated titleIJCB
CountryUnited States
CityClearwater
Period29/09/142/10/14

Keywords

  • SCS-Cybersecurity
  • recognition performance
  • EWI-25577
  • METIS-309819
  • Face Recognition
  • IR-93651
  • Image quality

Cite this

Dutta, A. ; Veldhuis, Raymond N.J. ; Spreeuwers, Lieuwe Jan. / A Bayesian model for predicting face recognition performance using image quality. 2014 IEEE International Joint Conference on Biometrics (IJCB). USA : IEEE, 2014. pp. 1-8
@inproceedings{fe77ae58b73e472ab8dbe6821c3335a9,
title = "A Bayesian model for predicting face recognition performance using image quality",
abstract = "Quality of a pair of facial images is a strong indicator of the uncertainty in decision about identity based on that image pair. In this paper, we describe a Bayesian approach to model the relation between image quality (like pose, illumination, noise, sharpness, etc) and corresponding face recognition performance. Experiment results based on the MultiPIE data set show that our model can accurately aggregate verification samples into groups for which the verification performance varies fairly consistently. Our model does not require similarity scores and can predict face recognition performance using only image quality information. Such a model has many applications. As an illustrative application, we show improved verification performance when the decision threshold automatically adapts according to the quality of facial images.",
keywords = "SCS-Cybersecurity, recognition performance, EWI-25577, METIS-309819, Face Recognition, IR-93651, Image quality",
author = "A. Dutta and Veldhuis, {Raymond N.J.} and Spreeuwers, {Lieuwe Jan}",
note = "eemcs-eprint-25577",
year = "2014",
month = "9",
day = "29",
doi = "10.1109/BTAS.2014.6996248",
language = "Undefined",
isbn = "978-1-4799-3584-0",
publisher = "IEEE",
pages = "1--8",
booktitle = "2014 IEEE International Joint Conference on Biometrics (IJCB)",
address = "United States",

}

Dutta, A, Veldhuis, RNJ & Spreeuwers, LJ 2014, A Bayesian model for predicting face recognition performance using image quality. in 2014 IEEE International Joint Conference on Biometrics (IJCB). IEEE, USA, pp. 1-8, 2014 IEEE International Joint Conference on Biometrics, IJCB 2014, Clearwater, United States, 29/09/14. https://doi.org/10.1109/BTAS.2014.6996248

A Bayesian model for predicting face recognition performance using image quality. / Dutta, A.; Veldhuis, Raymond N.J.; Spreeuwers, Lieuwe Jan.

2014 IEEE International Joint Conference on Biometrics (IJCB). USA : IEEE, 2014. p. 1-8.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - A Bayesian model for predicting face recognition performance using image quality

AU - Dutta, A.

AU - Veldhuis, Raymond N.J.

AU - Spreeuwers, Lieuwe Jan

N1 - eemcs-eprint-25577

PY - 2014/9/29

Y1 - 2014/9/29

N2 - Quality of a pair of facial images is a strong indicator of the uncertainty in decision about identity based on that image pair. In this paper, we describe a Bayesian approach to model the relation between image quality (like pose, illumination, noise, sharpness, etc) and corresponding face recognition performance. Experiment results based on the MultiPIE data set show that our model can accurately aggregate verification samples into groups for which the verification performance varies fairly consistently. Our model does not require similarity scores and can predict face recognition performance using only image quality information. Such a model has many applications. As an illustrative application, we show improved verification performance when the decision threshold automatically adapts according to the quality of facial images.

AB - Quality of a pair of facial images is a strong indicator of the uncertainty in decision about identity based on that image pair. In this paper, we describe a Bayesian approach to model the relation between image quality (like pose, illumination, noise, sharpness, etc) and corresponding face recognition performance. Experiment results based on the MultiPIE data set show that our model can accurately aggregate verification samples into groups for which the verification performance varies fairly consistently. Our model does not require similarity scores and can predict face recognition performance using only image quality information. Such a model has many applications. As an illustrative application, we show improved verification performance when the decision threshold automatically adapts according to the quality of facial images.

KW - SCS-Cybersecurity

KW - recognition performance

KW - EWI-25577

KW - METIS-309819

KW - Face Recognition

KW - IR-93651

KW - Image quality

U2 - 10.1109/BTAS.2014.6996248

DO - 10.1109/BTAS.2014.6996248

M3 - Conference contribution

SN - 978-1-4799-3584-0

SP - 1

EP - 8

BT - 2014 IEEE International Joint Conference on Biometrics (IJCB)

PB - IEEE

CY - USA

ER -

Dutta A, Veldhuis RNJ, Spreeuwers LJ. A Bayesian model for predicting face recognition performance using image quality. In 2014 IEEE International Joint Conference on Biometrics (IJCB). USA: IEEE. 2014. p. 1-8 https://doi.org/10.1109/BTAS.2014.6996248