File Download
  Links for fulltext
     (May Require Subscription)
Supplementary

Article: Robust speech recognition based on a Bayesian prediction approach

TitleRobust speech recognition based on a Bayesian prediction approach
Authors
Issue Date1999
PublisherIEEE.
Citation
IEEE Transactions on Speech and Audio Processing, 1999, v. 7 n. 4, p. 426-440 How to Cite?
AbstractWe study a category of robust speech recognition problem in which mismatches exist between training and testing conditions, and no accurate knowledge of the mismatch mechanism is available. The only available information is the test data along with a set of pretrained Gaussian mixture continuous density hidden Markov models (CDHMMs). We investigate the problem from the viewpoint of Bayesian prediction. A simple prior distribution, namely constrained uniform distribution, is adopted to characterize the uncertainty of the mean vectors of the CDHMMs. Two methods, namely a model compensation technique based on Bayesian predictive density and a robust decision strategy called Viterbi Bayesian predictive classification are studied. The proposed methods are compared with the conventional Viterbi decoding algorithm in speaker-independent recognition experiments on isolated digits and TI connected digit strings (TIDTGITS), where the mismatches between training and testing conditions are caused by: (1) additive Gaussian white noise, (2) each of 25 types of actual additive ambient noises, and (3) gender difference. The experimental results show that the adopted prior distribution and the proposed techniques help to improve the performance robustness under the examined mismatch conditions.
Persistent Identifierhttp://hdl.handle.net/10722/43648
ISSN
2007 Impact Factor: 2.291

 

DC FieldValueLanguage
dc.contributor.authorJiang, Hen_HK
dc.contributor.authorHirose, Ken_HK
dc.contributor.authorHuo, Qen_HK
dc.date.accessioned2007-03-23T04:51:13Z-
dc.date.available2007-03-23T04:51:13Z-
dc.date.issued1999en_HK
dc.identifier.citationIEEE Transactions on Speech and Audio Processing, 1999, v. 7 n. 4, p. 426-440en_HK
dc.identifier.issn1063-6676en_HK
dc.identifier.urihttp://hdl.handle.net/10722/43648-
dc.description.abstractWe study a category of robust speech recognition problem in which mismatches exist between training and testing conditions, and no accurate knowledge of the mismatch mechanism is available. The only available information is the test data along with a set of pretrained Gaussian mixture continuous density hidden Markov models (CDHMMs). We investigate the problem from the viewpoint of Bayesian prediction. A simple prior distribution, namely constrained uniform distribution, is adopted to characterize the uncertainty of the mean vectors of the CDHMMs. Two methods, namely a model compensation technique based on Bayesian predictive density and a robust decision strategy called Viterbi Bayesian predictive classification are studied. The proposed methods are compared with the conventional Viterbi decoding algorithm in speaker-independent recognition experiments on isolated digits and TI connected digit strings (TIDTGITS), where the mismatches between training and testing conditions are caused by: (1) additive Gaussian white noise, (2) each of 25 types of actual additive ambient noises, and (3) gender difference. The experimental results show that the adopted prior distribution and the proposed techniques help to improve the performance robustness under the examined mismatch conditions.en_HK
dc.format.extent464112 bytes-
dc.format.extent27136 bytes-
dc.format.mimetypeapplication/pdf-
dc.format.mimetypeapplication/msword-
dc.languageengen_HK
dc.publisherIEEE.en_HK
dc.rightsCreative Commons: Attribution 3.0 Hong Kong License-
dc.rights©1999 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.en_HK
dc.titleRobust speech recognition based on a Bayesian prediction approachen_HK
dc.typeArticleen_HK
dc.identifier.openurlhttp://library.hku.hk:4550/resserv?sid=HKU:IR&issn=1063-6676&volume=7&issue=4&spage=426&epage=440&date=1999&atitle=Robust+speech+recognition+based+on+a+Bayesian+prediction+approachen_HK
dc.description.naturepublished_or_final_versionen_HK
dc.identifier.doi10.1109/89.771309en_HK
dc.identifier.scopuseid_2-s2.0-0032685060-
dc.identifier.hkuros47887-

Export via OAI-PMH Interface in XML Formats


OR


Export to Other Non-XML Formats