Probabilistic random projections and speaker verification

Chong Lee Ying, Andrew Teoh Beng Jin

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

4 Citations (Scopus)

Abstract

Biometrics is susceptible to non-revocable and privacy invasion problems. Multiple Random Projections (MRP) was introduced as one of the cancellable biometrics approaches in face recognition to tackle these issues. However, this technique is applicable only to ID fixed length biometric feature vector but failed in varying size feature, such as speech biometrics. Besides, simple matching metric that used in MRP unable to offer a satisfactory verification performance. In this paper, we propose a variant of MRP, coined as Probabilistic Random Projections (PRP) in text-independent speaker verification. The PRP represents speech feature in 2D matrix format and speaker modeling is implemented through Gaussian Mixture Model. The formulation is experimented under two scenarios (legitimate and stolen token) using YOHO speech database. Besides that, desired properties such as one-way transformation and diversity are also examined.

Original languageEnglish
Title of host publicationAdvances in Biometrics - International Conference, ICB 2007, Proceedings
Pages445-454
Number of pages10
Publication statusPublished - 2007 Dec 1
Event2007 International Conference on Advances in Biometrics, ICB 2007 - Seoul, Korea, Republic of
Duration: 2007 Aug 272007 Aug 29

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4642 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other2007 International Conference on Advances in Biometrics, ICB 2007
CountryKorea, Republic of
CitySeoul
Period07/8/2707/8/29

Fingerprint

Random Projection
Speaker Verification
Biometrics
Face recognition
Gaussian Mixture Model
Invasion
Feature Vector
Face Recognition
Privacy
Metric
Scenarios
Formulation
Modeling
Speech

All Science Journal Classification (ASJC) codes

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Ying, C. L., & Jin, A. T. B. (2007). Probabilistic random projections and speaker verification. In Advances in Biometrics - International Conference, ICB 2007, Proceedings (pp. 445-454). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4642 LNCS).
Ying, Chong Lee ; Jin, Andrew Teoh Beng. / Probabilistic random projections and speaker verification. Advances in Biometrics - International Conference, ICB 2007, Proceedings. 2007. pp. 445-454 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{ba19e44d0aea454798777011fb65820b,
title = "Probabilistic random projections and speaker verification",
abstract = "Biometrics is susceptible to non-revocable and privacy invasion problems. Multiple Random Projections (MRP) was introduced as one of the cancellable biometrics approaches in face recognition to tackle these issues. However, this technique is applicable only to ID fixed length biometric feature vector but failed in varying size feature, such as speech biometrics. Besides, simple matching metric that used in MRP unable to offer a satisfactory verification performance. In this paper, we propose a variant of MRP, coined as Probabilistic Random Projections (PRP) in text-independent speaker verification. The PRP represents speech feature in 2D matrix format and speaker modeling is implemented through Gaussian Mixture Model. The formulation is experimented under two scenarios (legitimate and stolen token) using YOHO speech database. Besides that, desired properties such as one-way transformation and diversity are also examined.",
author = "Ying, {Chong Lee} and Jin, {Andrew Teoh Beng}",
year = "2007",
month = "12",
day = "1",
language = "English",
isbn = "9783540745488",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "445--454",
booktitle = "Advances in Biometrics - International Conference, ICB 2007, Proceedings",

}

Ying, CL & Jin, ATB 2007, Probabilistic random projections and speaker verification. in Advances in Biometrics - International Conference, ICB 2007, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 4642 LNCS, pp. 445-454, 2007 International Conference on Advances in Biometrics, ICB 2007, Seoul, Korea, Republic of, 07/8/27.

Probabilistic random projections and speaker verification. / Ying, Chong Lee; Jin, Andrew Teoh Beng.

Advances in Biometrics - International Conference, ICB 2007, Proceedings. 2007. p. 445-454 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 4642 LNCS).

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

TY - GEN

T1 - Probabilistic random projections and speaker verification

AU - Ying, Chong Lee

AU - Jin, Andrew Teoh Beng

PY - 2007/12/1

Y1 - 2007/12/1

N2 - Biometrics is susceptible to non-revocable and privacy invasion problems. Multiple Random Projections (MRP) was introduced as one of the cancellable biometrics approaches in face recognition to tackle these issues. However, this technique is applicable only to ID fixed length biometric feature vector but failed in varying size feature, such as speech biometrics. Besides, simple matching metric that used in MRP unable to offer a satisfactory verification performance. In this paper, we propose a variant of MRP, coined as Probabilistic Random Projections (PRP) in text-independent speaker verification. The PRP represents speech feature in 2D matrix format and speaker modeling is implemented through Gaussian Mixture Model. The formulation is experimented under two scenarios (legitimate and stolen token) using YOHO speech database. Besides that, desired properties such as one-way transformation and diversity are also examined.

AB - Biometrics is susceptible to non-revocable and privacy invasion problems. Multiple Random Projections (MRP) was introduced as one of the cancellable biometrics approaches in face recognition to tackle these issues. However, this technique is applicable only to ID fixed length biometric feature vector but failed in varying size feature, such as speech biometrics. Besides, simple matching metric that used in MRP unable to offer a satisfactory verification performance. In this paper, we propose a variant of MRP, coined as Probabilistic Random Projections (PRP) in text-independent speaker verification. The PRP represents speech feature in 2D matrix format and speaker modeling is implemented through Gaussian Mixture Model. The formulation is experimented under two scenarios (legitimate and stolen token) using YOHO speech database. Besides that, desired properties such as one-way transformation and diversity are also examined.

UR - http://www.scopus.com/inward/record.url?scp=37849030222&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=37849030222&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:37849030222

SN - 9783540745488

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 445

EP - 454

BT - Advances in Biometrics - International Conference, ICB 2007, Proceedings

ER -

Ying CL, Jin ATB. Probabilistic random projections and speaker verification. In Advances in Biometrics - International Conference, ICB 2007, Proceedings. 2007. p. 445-454. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).