MAP estimation of subspace transform for speaker recognition

Donglai Zhu; Bin Ma; Kong Aik Lee; Cheung Chi Leung; Haizhou Li

MAP estimation of subspace transform for speaker recognition

Donglai Zhu, Bin Ma, Kong Aik Lee, Cheung Chi Leung, Haizhou Li

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

Abstract

We propose using the maximum-a-posteriori (MAP) estimation of subspace transform for speaker recognition. The linear transform is defined on the mean vectors of the Gaussian mixture model (GMM), where transform matrices and bias vectors are associated with separate regression classes so that both can be estimated with sufficient statistics given limited training data. The transform matrices are further defined as a linear combination of a set of basis transforms so that the weights are parameters to be estimated. We characterize the speakers with the transform parameters and model them using support vector machine (SVM). Experiments on the 2008 NIST SRE task illustrate the effectiveness of the method.

Original language	English
Title of host publication	Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010
Publisher	International Speech Communication Association
Pages	1465-1468
Number of pages	4
Publication status	Published - Sept 2010
Externally published	Yes

Publication series

Name	Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010

Keywords

Maximum a posteriori
Speaker recognition
Subspace transform

ASJC Scopus subject areas

Language and Linguistics
Speech and Hearing
Human-Computer Interaction
Signal Processing
Software
Modelling and Simulation

Cite this

Zhu, D., Ma, B., Lee, K. A., Leung, C. C., & Li, H. (2010). MAP estimation of subspace transform for speaker recognition. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010 (pp. 1465-1468). (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010). International Speech Communication Association.

Zhu, Donglai ; Ma, Bin ; Lee, Kong Aik et al. / MAP estimation of subspace transform for speaker recognition. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. pp. 1465-1468 (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

@inproceedings{d8d7578e4cf7446eb772c92a296d1fb1,

title = "MAP estimation of subspace transform for speaker recognition",

abstract = "We propose using the maximum-a-posteriori (MAP) estimation of subspace transform for speaker recognition. The linear transform is defined on the mean vectors of the Gaussian mixture model (GMM), where transform matrices and bias vectors are associated with separate regression classes so that both can be estimated with sufficient statistics given limited training data. The transform matrices are further defined as a linear combination of a set of basis transforms so that the weights are parameters to be estimated. We characterize the speakers with the transform parameters and model them using support vector machine (SVM). Experiments on the 2008 NIST SRE task illustrate the effectiveness of the method.",

keywords = "Maximum a posteriori, Speaker recognition, Subspace transform",

author = "Donglai Zhu and Bin Ma and Lee, {Kong Aik} and Leung, {Cheung Chi} and Haizhou Li",

year = "2010",

month = sep,

language = "English",

series = "Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010",

publisher = "International Speech Communication Association",

pages = "1465--1468",

booktitle = "Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010",

}

Zhu, D, Ma, B, Lee, KA, Leung, CC & Li, H 2010, MAP estimation of subspace transform for speaker recognition. in Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010, International Speech Communication Association, pp. 1465-1468.

MAP estimation of subspace transform for speaker recognition. / Zhu, Donglai; Ma, Bin; Lee, Kong Aik et al.
Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. p. 1465-1468 (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

Research output: Chapter in book / Conference proceeding › Conference article published in proceeding or book › Academic research › peer-review

TY - GEN

T1 - MAP estimation of subspace transform for speaker recognition

AU - Zhu, Donglai

AU - Ma, Bin

AU - Lee, Kong Aik

AU - Leung, Cheung Chi

AU - Li, Haizhou

PY - 2010/9

Y1 - 2010/9

N2 - We propose using the maximum-a-posteriori (MAP) estimation of subspace transform for speaker recognition. The linear transform is defined on the mean vectors of the Gaussian mixture model (GMM), where transform matrices and bias vectors are associated with separate regression classes so that both can be estimated with sufficient statistics given limited training data. The transform matrices are further defined as a linear combination of a set of basis transforms so that the weights are parameters to be estimated. We characterize the speakers with the transform parameters and model them using support vector machine (SVM). Experiments on the 2008 NIST SRE task illustrate the effectiveness of the method.

AB - We propose using the maximum-a-posteriori (MAP) estimation of subspace transform for speaker recognition. The linear transform is defined on the mean vectors of the Gaussian mixture model (GMM), where transform matrices and bias vectors are associated with separate regression classes so that both can be estimated with sufficient statistics given limited training data. The transform matrices are further defined as a linear combination of a set of basis transforms so that the weights are parameters to be estimated. We characterize the speakers with the transform parameters and model them using support vector machine (SVM). Experiments on the 2008 NIST SRE task illustrate the effectiveness of the method.

KW - Maximum a posteriori

KW - Speaker recognition

KW - Subspace transform

UR - http://www.scopus.com/inward/record.url?scp=79959837604&partnerID=8YFLogxK

M3 - Conference article published in proceeding or book

AN - SCOPUS:79959837604

T3 - Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010

SP - 1465

EP - 1468

BT - Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010

PB - International Speech Communication Association

ER -

Zhu D, Ma B, Lee KA, Leung CC, Li H. MAP estimation of subspace transform for speaker recognition. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association. 2010. p. 1465-1468. (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

MAP estimation of subspace transform for speaker recognition

Abstract

Publication series

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this