Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation

Yongxin Zhang, Michael S Scordilis

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

A new statistical Voiced/Unvoiced/Mixed/Silence classifier based on a Maximum a Posteriori (MAP) adaptation algorithm is presented. The speech signal distributions are modeled with Gaussian Mixture Models (GMM). The MAP re-estimation of model parameters is based on the sufficient statistics within a form of Bayesian adaptation. The robustness of the proposed technique and model adaptation to different background/channel conditions were evaluated. Experimental results show that the proposed method can adapt and is robust to adverse signal conditions, such as SNR as low as 3dB, noise in a moving vehicle, and band-limited channel conditions.

Original languageEnglish (US)
Title of host publicationConference Proceedings - IEEE SOUTHEASTCON
EditorsY. Levy
Pages229-232
Number of pages4
StatePublished - 2005
EventIEEE Southeastcon 2005: Excellence in Engineering, Science and Technology - Ft. Lauderdale, United Kingdom
Duration: Apr 8 2005Apr 10 2005

Other

OtherIEEE Southeastcon 2005: Excellence in Engineering, Science and Technology
CountryUnited Kingdom
CityFt. Lauderdale
Period4/8/054/10/05

Fingerprint

Classifiers
Statistics

ASJC Scopus subject areas

  • Electrical and Electronic Engineering

Cite this

Zhang, Y., & Scordilis, M. S. (2005). Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation. In Y. Levy (Ed.), Conference Proceedings - IEEE SOUTHEASTCON (pp. 229-232)

Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation. / Zhang, Yongxin; Scordilis, Michael S.

Conference Proceedings - IEEE SOUTHEASTCON. ed. / Y. Levy. 2005. p. 229-232.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhang, Y & Scordilis, MS 2005, Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation. in Y Levy (ed.), Conference Proceedings - IEEE SOUTHEASTCON. pp. 229-232, IEEE Southeastcon 2005: Excellence in Engineering, Science and Technology, Ft. Lauderdale, United Kingdom, 4/8/05.
Zhang Y, Scordilis MS. Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation. In Levy Y, editor, Conference Proceedings - IEEE SOUTHEASTCON. 2005. p. 229-232
Zhang, Yongxin ; Scordilis, Michael S. / Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation. Conference Proceedings - IEEE SOUTHEASTCON. editor / Y. Levy. 2005. pp. 229-232
@inproceedings{5bae9be380ea45b6bd42af7a0f35d75c,
title = "Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation",
abstract = "A new statistical Voiced/Unvoiced/Mixed/Silence classifier based on a Maximum a Posteriori (MAP) adaptation algorithm is presented. The speech signal distributions are modeled with Gaussian Mixture Models (GMM). The MAP re-estimation of model parameters is based on the sufficient statistics within a form of Bayesian adaptation. The robustness of the proposed technique and model adaptation to different background/channel conditions were evaluated. Experimental results show that the proposed method can adapt and is robust to adverse signal conditions, such as SNR as low as 3dB, noise in a moving vehicle, and band-limited channel conditions.",
author = "Yongxin Zhang and Scordilis, {Michael S}",
year = "2005",
language = "English (US)",
pages = "229--232",
editor = "Y. Levy",
booktitle = "Conference Proceedings - IEEE SOUTHEASTCON",

}

TY - GEN

T1 - Robust voiced/unvoiced/mixed/silence classifier with maximum a Posteriori channel/background adaptation

AU - Zhang, Yongxin

AU - Scordilis, Michael S

PY - 2005

Y1 - 2005

N2 - A new statistical Voiced/Unvoiced/Mixed/Silence classifier based on a Maximum a Posteriori (MAP) adaptation algorithm is presented. The speech signal distributions are modeled with Gaussian Mixture Models (GMM). The MAP re-estimation of model parameters is based on the sufficient statistics within a form of Bayesian adaptation. The robustness of the proposed technique and model adaptation to different background/channel conditions were evaluated. Experimental results show that the proposed method can adapt and is robust to adverse signal conditions, such as SNR as low as 3dB, noise in a moving vehicle, and band-limited channel conditions.

AB - A new statistical Voiced/Unvoiced/Mixed/Silence classifier based on a Maximum a Posteriori (MAP) adaptation algorithm is presented. The speech signal distributions are modeled with Gaussian Mixture Models (GMM). The MAP re-estimation of model parameters is based on the sufficient statistics within a form of Bayesian adaptation. The robustness of the proposed technique and model adaptation to different background/channel conditions were evaluated. Experimental results show that the proposed method can adapt and is robust to adverse signal conditions, such as SNR as low as 3dB, noise in a moving vehicle, and band-limited channel conditions.

UR - http://www.scopus.com/inward/record.url?scp=27544437061&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=27544437061&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:27544437061

SP - 229

EP - 232

BT - Conference Proceedings - IEEE SOUTHEASTCON

A2 - Levy, Y.

ER -