Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification

Yi Liu; Liang He; Wei Qiang Zhang; Jia Liu; Michael T. Johnson

doi:10.23919/APSIPA.2018.8659790

Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification

Yi Liu, Liang He, Wei Qiang Zhang, Jia Liu, Michael T. Johnson

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

2 Scopus citations

Abstract

Frame alignments can be computed by different methods in GMM-based speaker verification. By incorporating a phonetic Gaussian mixture model (PGMM), we are able to compare the performance using alignments extracted from the deep neural networks (DNN) and the conventional hidden Markov model (HMM) in digit-prompted speaker verification. Based on the different characteristics of these two alignments, we present a novel content verification method to improve the system security without much computational overhead. Our experiments on the RSR2015 Part-3 digit-prompted task show that, the DNN-based alignment performs on par with the HMM alignment. The results also demonstrate the effectiveness of the proposed Kullback-Leibler (KL) divergence based scoring to reject speech with incorrect pass-phrases.

Original language	English
Title of host publication	2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings
Pages	1467-1472
Number of pages	6
ISBN (Electronic)	9789881476852
DOIs	https://doi.org/10.23919/APSIPA.2018.8659790
State	Published - Jul 2 2018
Event	10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Honolulu, United States Duration: Nov 12 2018 → Nov 15 2018

Publication series

Name	2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings

Conference

Conference	10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018
Country/Territory	United States
City	Honolulu
Period	11/12/18 → 11/15/18

Bibliographical note

Publisher Copyright:
© 2018 APSIPA organization.

Funding

The work is supported by National Natural Science Foundation of China under Grant No. 61370034, No. 61403224 and No. 61273268.

Funders	Funder number
National Natural Science Foundation of P.R. China	61403224, 61273268, 61370034
National Natural Science Foundation of P.R. China

ASJC Scopus subject areas

Information Systems

Access to Document

10.23919/APSIPA.2018.8659790

Cite this

Liu, Y., He, L., Zhang, W. Q., Liu, J., & Johnson, M. T. (2018). Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification. In 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings (pp. 1467-1472). Article 8659790 (2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings). https://doi.org/10.23919/APSIPA.2018.8659790

Liu, Yi ; He, Liang ; Zhang, Wei Qiang et al. / Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification. 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings. 2018. pp. 1467-1472 (2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings).

@inproceedings{aeee6de987a749b0a07b9587a118cc10,

title = "Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification",

abstract = "Frame alignments can be computed by different methods in GMM-based speaker verification. By incorporating a phonetic Gaussian mixture model (PGMM), we are able to compare the performance using alignments extracted from the deep neural networks (DNN) and the conventional hidden Markov model (HMM) in digit-prompted speaker verification. Based on the different characteristics of these two alignments, we present a novel content verification method to improve the system security without much computational overhead. Our experiments on the RSR2015 Part-3 digit-prompted task show that, the DNN-based alignment performs on par with the HMM alignment. The results also demonstrate the effectiveness of the proposed Kullback-Leibler (KL) divergence based scoring to reject speech with incorrect pass-phrases.",

author = "Yi Liu and Liang He and Zhang, \{Wei Qiang\} and Jia Liu and Johnson, \{Michael T.\}",

note = "Publisher Copyright: {\textcopyright} 2018 APSIPA organization.; 10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 ; Conference date: 12-11-2018 Through 15-11-2018",

year = "2018",

month = jul,

day = "2",

doi = "10.23919/APSIPA.2018.8659790",

language = "English",

series = "2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings",

pages = "1467--1472",

booktitle = "2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings",

}

Liu, Y, He, L, Zhang, WQ, Liu, J & Johnson, MT 2018, Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification. in 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings., 8659790, 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings, pp. 1467-1472, 10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018, Honolulu, United States, 11/12/18. https://doi.org/10.23919/APSIPA.2018.8659790

Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification. / Liu, Yi; He, Liang; Zhang, Wei Qiang et al.
2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings. 2018. p. 1467-1472 8659790 (2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings).

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

TY - GEN

T1 - Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification

AU - Liu, Yi

AU - He, Liang

AU - Zhang, Wei Qiang

AU - Liu, Jia

AU - Johnson, Michael T.

PY - 2018/7/2

Y1 - 2018/7/2

N2 - Frame alignments can be computed by different methods in GMM-based speaker verification. By incorporating a phonetic Gaussian mixture model (PGMM), we are able to compare the performance using alignments extracted from the deep neural networks (DNN) and the conventional hidden Markov model (HMM) in digit-prompted speaker verification. Based on the different characteristics of these two alignments, we present a novel content verification method to improve the system security without much computational overhead. Our experiments on the RSR2015 Part-3 digit-prompted task show that, the DNN-based alignment performs on par with the HMM alignment. The results also demonstrate the effectiveness of the proposed Kullback-Leibler (KL) divergence based scoring to reject speech with incorrect pass-phrases.

AB - Frame alignments can be computed by different methods in GMM-based speaker verification. By incorporating a phonetic Gaussian mixture model (PGMM), we are able to compare the performance using alignments extracted from the deep neural networks (DNN) and the conventional hidden Markov model (HMM) in digit-prompted speaker verification. Based on the different characteristics of these two alignments, we present a novel content verification method to improve the system security without much computational overhead. Our experiments on the RSR2015 Part-3 digit-prompted task show that, the DNN-based alignment performs on par with the HMM alignment. The results also demonstrate the effectiveness of the proposed Kullback-Leibler (KL) divergence based scoring to reject speech with incorrect pass-phrases.

UR - http://www.scopus.com/inward/record.url?scp=85063451765&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063451765&partnerID=8YFLogxK

U2 - 10.23919/APSIPA.2018.8659790

DO - 10.23919/APSIPA.2018.8659790

M3 - Conference contribution

AN - SCOPUS:85063451765

T3 - 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings

SP - 1467

EP - 1472

BT - 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings

T2 - 10th Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018

Y2 - 12 November 2018 through 15 November 2018

ER -

Liu Y, He L, Zhang WQ, Liu J, Johnson MT. Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification. In 2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings. 2018. p. 1467-1472. 8659790. (2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference, APSIPA ASC 2018 - Proceedings). doi: 10.23919/APSIPA.2018.8659790

Investigation of Frame Alignments for GMM-based Digit-prompted Speaker Verification

Abstract

Publication series

Conference

Bibliographical note

Funding

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this