Robust analysis and weighting on MFCC components for speech recognition and speaker identification

Xi Zhou; Yun Fu; Ming Liu; Mark Hasegawa-Johnson; Thomas S. Huang

doi:10.1109/icme.2007.4284618

Robust analysis and weighting on MFCC components for speech recognition and speaker identification

Xi Zhou, Yun Fu, Ming Liu, Mark Hasegawa-Johnson, Thomas S. Huang

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

Abstract

Mismatch between training and testing data is a major error source for both Automatic Speech Recognition (ASR) and Automatic Speaker Identification (ASI). In this paper, we first present a statistical weighting concept to exploit the unequal sensitivity of Mel-Frequency Cepstral Coefficients (MFCC) components to against the mismatch, such as ambient noise, recording equipment, transmission channels, and inter-speaker variations. We further design a new Kullback-Leibler (KL) Distance based weighting algorithm according to the proposed weighting concept to real-world problems in which the label information is often not provided. We examine our algorithm in ASR with mismatch by different speakers and also in ASI with mismatch by channel noises. Experimental results demonstrate the effectiveness and robustness of our proposed method.

Original language	English (US)
Title of host publication	Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007
Publisher	IEEE Computer Society
Pages	188-191
Number of pages	4
ISBN (Print)	1424410177, 9781424410170
DOIs	https://doi.org/10.1109/icme.2007.4284618
State	Published - 2007
Event	IEEE International Conference onMultimedia and Expo, ICME 2007 - Beijing, China Duration: Jul 2 2007 → Jul 5 2007

Publication series

Name	Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007

Other

Other	IEEE International Conference onMultimedia and Expo, ICME 2007
Country/Territory	China
City	Beijing
Period	7/2/07 → 7/5/07

ASJC Scopus subject areas

Computer Graphics and Computer-Aided Design
Software

Online availability

10.1109/icme.2007.4284618

Library availability

Discover UIUC Full Text

Cite this

Zhou, X., Fu, Y., Liu, M., Hasegawa-Johnson, M., & Huang, T. S. (2007). Robust analysis and weighting on MFCC components for speech recognition and speaker identification. In Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007 (pp. 188-191). Article 4284618 (Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007). IEEE Computer Society. https://doi.org/10.1109/icme.2007.4284618

Robust analysis and weighting on MFCC components for speech recognition and speaker identification. / Zhou, Xi; Fu, Yun; Liu, Ming et al.
Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007. IEEE Computer Society, 2007. p. 188-191 4284618 (Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007).

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

Zhou, X, Fu, Y, Liu, M, Hasegawa-Johnson, M & Huang, TS 2007, Robust analysis and weighting on MFCC components for speech recognition and speaker identification. in Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007., 4284618, Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007, IEEE Computer Society, pp. 188-191, IEEE International Conference onMultimedia and Expo, ICME 2007, Beijing, China, 7/2/07. https://doi.org/10.1109/icme.2007.4284618

Zhou X, Fu Y, Liu M, Hasegawa-Johnson M, Huang TS. Robust analysis and weighting on MFCC components for speech recognition and speaker identification. In Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007. IEEE Computer Society. 2007. p. 188-191. 4284618. (Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007). doi: 10.1109/icme.2007.4284618

@inproceedings{e41356a5bab64c59af451eb29467c9b0,

title = "Robust analysis and weighting on MFCC components for speech recognition and speaker identification",

abstract = "Mismatch between training and testing data is a major error source for both Automatic Speech Recognition (ASR) and Automatic Speaker Identification (ASI). In this paper, we first present a statistical weighting concept to exploit the unequal sensitivity of Mel-Frequency Cepstral Coefficients (MFCC) components to against the mismatch, such as ambient noise, recording equipment, transmission channels, and inter-speaker variations. We further design a new Kullback-Leibler (KL) Distance based weighting algorithm according to the proposed weighting concept to real-world problems in which the label information is often not provided. We examine our algorithm in ASR with mismatch by different speakers and also in ASI with mismatch by channel noises. Experimental results demonstrate the effectiveness and robustness of our proposed method.",

author = "Xi Zhou and Yun Fu and Ming Liu and Mark Hasegawa-Johnson and Huang, {Thomas S.}",

year = "2007",

doi = "10.1109/icme.2007.4284618",

language = "English (US)",

isbn = "1424410177",

series = "Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007",

publisher = "IEEE Computer Society",

pages = "188--191",

booktitle = "Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007",

note = "IEEE International Conference onMultimedia and Expo, ICME 2007 ; Conference date: 02-07-2007 Through 05-07-2007",

}

TY - GEN

T1 - Robust analysis and weighting on MFCC components for speech recognition and speaker identification

AU - Zhou, Xi

AU - Fu, Yun

AU - Liu, Ming

AU - Hasegawa-Johnson, Mark

AU - Huang, Thomas S.

PY - 2007

Y1 - 2007

N2 - Mismatch between training and testing data is a major error source for both Automatic Speech Recognition (ASR) and Automatic Speaker Identification (ASI). In this paper, we first present a statistical weighting concept to exploit the unequal sensitivity of Mel-Frequency Cepstral Coefficients (MFCC) components to against the mismatch, such as ambient noise, recording equipment, transmission channels, and inter-speaker variations. We further design a new Kullback-Leibler (KL) Distance based weighting algorithm according to the proposed weighting concept to real-world problems in which the label information is often not provided. We examine our algorithm in ASR with mismatch by different speakers and also in ASI with mismatch by channel noises. Experimental results demonstrate the effectiveness and robustness of our proposed method.

AB - Mismatch between training and testing data is a major error source for both Automatic Speech Recognition (ASR) and Automatic Speaker Identification (ASI). In this paper, we first present a statistical weighting concept to exploit the unequal sensitivity of Mel-Frequency Cepstral Coefficients (MFCC) components to against the mismatch, such as ambient noise, recording equipment, transmission channels, and inter-speaker variations. We further design a new Kullback-Leibler (KL) Distance based weighting algorithm according to the proposed weighting concept to real-world problems in which the label information is often not provided. We examine our algorithm in ASR with mismatch by different speakers and also in ASI with mismatch by channel noises. Experimental results demonstrate the effectiveness and robustness of our proposed method.

UR - http://www.scopus.com/inward/record.url?scp=46449092074&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=46449092074&partnerID=8YFLogxK

U2 - 10.1109/icme.2007.4284618

DO - 10.1109/icme.2007.4284618

M3 - Conference contribution

AN - SCOPUS:46449092074

SN - 1424410177

SN - 9781424410170

T3 - Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007

SP - 188

EP - 191

BT - Proceedings of the 2007 IEEE International Conference on Multimedia and Expo, ICME 2007

PB - IEEE Computer Society

T2 - IEEE International Conference onMultimedia and Expo, ICME 2007

Y2 - 2 July 2007 through 5 July 2007

ER -

Robust analysis and weighting on MFCC components for speech recognition and speaker identification

Abstract

Publication series

Other

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this