A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification

Chi Sang Jung; Kyu J. Han; Hyunson Seo; Shrikanth S. Narayanan; Hong Goo Kang

A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification

Chi Sang Jung, Kyu J. Han, Hyunson Seo, Shrikanth S. Narayanan, Hong Goo Kang

Department of Electrical and Electronic Engineering

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

9 Citations (Scopus)

Abstract

In this paper, we propose a spectral kurtosis based approach to extract features with a variable frame length and rate for speaker verification. Since the speaker-specific information of features in each frame changes depending upon the characteristics of speech, it is important to determine the appropriate frame length and rate to extract the salient feature frames. In order to distinctively represent the characteristics of vowels and consonants both in time and frequency domains, we introduce a variable frame length and rate (VFLR) method based on spectral kurtosis, which provides a local measure of time-frequency concentration. Experimental results verify that the proposed VFLR method improves the performance of the speaker verification system on the NIST SRE-06 database by 9.725% (relative) compared to the feature extraction method with the fixed length and rate.

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	2754-2757
Number of pages	4
Publication status	Published - 2010

Publication series

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

All Science Journal Classification (ASJC) codes

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

Cite this

Jung, C. S., Han, K. J., Seo, H., Narayanan, S. S., & Kang, H. G. (2010). A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010 (pp. 2754-2757). (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010). International Speech Communication Association.

Jung, Chi Sang ; Han, Kyu J. ; Seo, Hyunson et al. / A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification. Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. pp. 2754-2757 (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

@inproceedings{cd5d5073ca4541848e97deff5490bc9e,

title = "A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification",

abstract = "In this paper, we propose a spectral kurtosis based approach to extract features with a variable frame length and rate for speaker verification. Since the speaker-specific information of features in each frame changes depending upon the characteristics of speech, it is important to determine the appropriate frame length and rate to extract the salient feature frames. In order to distinctively represent the characteristics of vowels and consonants both in time and frequency domains, we introduce a variable frame length and rate (VFLR) method based on spectral kurtosis, which provides a local measure of time-frequency concentration. Experimental results verify that the proposed VFLR method improves the performance of the speaker verification system on the NIST SRE-06 database by 9.725% (relative) compared to the feature extraction method with the fixed length and rate.",

author = "Jung, {Chi Sang} and Han, {Kyu J.} and Hyunson Seo and Narayanan, {Shrikanth S.} and Kang, {Hong Goo}",

year = "2010",

language = "English",

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

publisher = "International Speech Communication Association",

pages = "2754--2757",

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

}

Jung, CS, Han, KJ, Seo, H, Narayanan, SS & Kang, HG 2010, A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification. 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. 2754-2757.

A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification. / Jung, Chi Sang; Han, Kyu J.; Seo, Hyunson et al.
Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association, 2010. p. 2754-2757 (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

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

TY - GEN

T1 - A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification

AU - Jung, Chi Sang

AU - Han, Kyu J.

AU - Seo, Hyunson

AU - Narayanan, Shrikanth S.

AU - Kang, Hong Goo

PY - 2010

Y1 - 2010

N2 - In this paper, we propose a spectral kurtosis based approach to extract features with a variable frame length and rate for speaker verification. Since the speaker-specific information of features in each frame changes depending upon the characteristics of speech, it is important to determine the appropriate frame length and rate to extract the salient feature frames. In order to distinctively represent the characteristics of vowels and consonants both in time and frequency domains, we introduce a variable frame length and rate (VFLR) method based on spectral kurtosis, which provides a local measure of time-frequency concentration. Experimental results verify that the proposed VFLR method improves the performance of the speaker verification system on the NIST SRE-06 database by 9.725% (relative) compared to the feature extraction method with the fixed length and rate.

AB - In this paper, we propose a spectral kurtosis based approach to extract features with a variable frame length and rate for speaker verification. Since the speaker-specific information of features in each frame changes depending upon the characteristics of speech, it is important to determine the appropriate frame length and rate to extract the salient feature frames. In order to distinctively represent the characteristics of vowels and consonants both in time and frequency domains, we introduce a variable frame length and rate (VFLR) method based on spectral kurtosis, which provides a local measure of time-frequency concentration. Experimental results verify that the proposed VFLR method improves the performance of the speaker verification system on the NIST SRE-06 database by 9.725% (relative) compared to the feature extraction method with the fixed length and rate.

UR - http://www.scopus.com/inward/record.url?scp=79959823356&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79959823356&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:79959823356

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

SP - 2754

EP - 2757

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

PB - International Speech Communication Association

ER -

Jung CS, Han KJ, Seo H, Narayanan SS, Kang HG. A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification. In Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010. International Speech Communication Association. 2010. p. 2754-2757. (Proceedings of the 11th Annual Conference of the International Speech Communication Association, INTERSPEECH 2010).

A variable frame length and rate algorithm based on the spectral kurtosis measure for speaker verification

Abstract

Publication series

All Science Journal Classification (ASJC) codes

Other files and links

Fingerprint

Cite this