A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor

Chang Huai You; Haizhou Li; Kong Aik Lee

A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor

Chang Huai You, Haizhou Li, Kong Aik Lee

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

3 Citations (Scopus)

Abstract

In Gaussian mixture model (GMM) approach to speaker recognition, it has been found that the maximum a posteriori (MAP) estimation is greatly affected by undesired variability due to varying duration of utterance as well as other hidden factors related to recording devices, session environment, and phonetic contents. We propose an adaptive relevance factor (RF) to compensate for this variability. In the other side, in realistic application, it is likely that the different channel corresponds to its different training and test conditions in terms of quantity and quality of the speech signals. In this connection, we develop a hybrid model that combines multiple complementary systems, each of which focuses on specific condition(s). We show the effectiveness of the proposed method on the core task of the National Institute of Standards and Technology (NIST) speaker recognition evaluation (SRE) 2008.

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	2746-2749
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

Gaussian mixture model
Maximum a posteriori
Speaker recognition

ASJC Scopus subject areas

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

Cite this

You, C. H., Li, H., & Lee, K. A. (2010). A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010 (pp. 2746-2749). (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010). International Speech Communication Association.

You, Chang Huai ; Li, Haizhou ; Lee, Kong Aik. / A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. pp. 2746-2749 (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

@inproceedings{24427e63acfa43a89fd0fe04adb83b94,

title = "A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor",

abstract = "In Gaussian mixture model (GMM) approach to speaker recognition, it has been found that the maximum a posteriori (MAP) estimation is greatly affected by undesired variability due to varying duration of utterance as well as other hidden factors related to recording devices, session environment, and phonetic contents. We propose an adaptive relevance factor (RF) to compensate for this variability. In the other side, in realistic application, it is likely that the different channel corresponds to its different training and test conditions in terms of quantity and quality of the speech signals. In this connection, we develop a hybrid model that combines multiple complementary systems, each of which focuses on specific condition(s). We show the effectiveness of the proposed method on the core task of the National Institute of Standards and Technology (NIST) speaker recognition evaluation (SRE) 2008.",

keywords = "Gaussian mixture model, Maximum a posteriori, Speaker recognition",

author = "You, {Chang Huai} and Haizhou Li and Lee, {Kong Aik}",

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 = "2746--2749",

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

}

You, CH, Li, H & Lee, KA 2010, A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor. 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. 2746-2749.

A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor. / You, Chang Huai; Li, Haizhou; Lee, Kong Aik.
Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. p. 2746-2749 (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 - A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor

AU - You, Chang Huai

AU - Li, Haizhou

AU - Lee, Kong Aik

PY - 2010/9

Y1 - 2010/9

N2 - In Gaussian mixture model (GMM) approach to speaker recognition, it has been found that the maximum a posteriori (MAP) estimation is greatly affected by undesired variability due to varying duration of utterance as well as other hidden factors related to recording devices, session environment, and phonetic contents. We propose an adaptive relevance factor (RF) to compensate for this variability. In the other side, in realistic application, it is likely that the different channel corresponds to its different training and test conditions in terms of quantity and quality of the speech signals. In this connection, we develop a hybrid model that combines multiple complementary systems, each of which focuses on specific condition(s). We show the effectiveness of the proposed method on the core task of the National Institute of Standards and Technology (NIST) speaker recognition evaluation (SRE) 2008.

AB - In Gaussian mixture model (GMM) approach to speaker recognition, it has been found that the maximum a posteriori (MAP) estimation is greatly affected by undesired variability due to varying duration of utterance as well as other hidden factors related to recording devices, session environment, and phonetic contents. We propose an adaptive relevance factor (RF) to compensate for this variability. In the other side, in realistic application, it is likely that the different channel corresponds to its different training and test conditions in terms of quantity and quality of the speech signals. In this connection, we develop a hybrid model that combines multiple complementary systems, each of which focuses on specific condition(s). We show the effectiveness of the proposed method on the core task of the National Institute of Standards and Technology (NIST) speaker recognition evaluation (SRE) 2008.

KW - Gaussian mixture model

KW - Maximum a posteriori

KW - Speaker recognition

UR - http://www.scopus.com/inward/record.url?scp=79959834641&partnerID=8YFLogxK

M3 - Conference article published in proceeding or book

AN - SCOPUS:79959834641

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

SP - 2746

EP - 2749

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

PB - International Speech Communication Association

ER -

You CH, Li H, Lee KA. A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association. 2010. p. 2746-2749. (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

A hybrid modeling strategy for GMM-SVM speaker recognition with adaptive relevance factor

Abstract

Publication series

Keywords

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this