InforLorV4, PascalFrancis, Corpus, bibRecord, 000C14

A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition

Identifieur interne : 000C14 ( PascalFrancis/Corpus ); précédent : 000C13; suivant : 000C15

A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition

Auteurs : M. Afify ; Y. Gong ; J.-P. Haton

Source :

RBID : Pascal:98-0082849

Descripteurs français

Pascal (Inist)
- Traitement signal, Rapport signal bruit, Processus Markov, Processus gaussien, Reconnaissance parole, Maximum vraisemblance, Erreur systématique, 4360, 4372.

English descriptors

KwdEn :
- Bias, Gaussian processes, Markov process, Maximum likelihood, Signal processing, Signal-to-noise ratio, Speech recognition.

Abstract

In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.

Notice en format standard (ISO 2709)

Pour connaître la documentation sur le format Inist Standard.

A08	`01`	`1`	`ENG`	`@1 A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition`
A09	`01`	`1`	`ENG`	`@1 ICASSP 97 : international conference on acoustics, speech, and signal processing : Munich, April 21-24, 1997. Volume II: Speech processing`
A11	`01`	`1`		`@1 AFIFY (M.)`
A11	`02`	`1`		`@1 GONG (Y.)`
A11	`03`	`1`		`@1 HATON (J.-P.)`
A14	`01`			`@1 CRIN/CNRS-INRIA-Lorraine, B.P. 239 @2 54506 Vandeouvre, Nancy @3 FRA @Z 1 aut. @Z 3 aut.`
A14	`02`			`@1 Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374 @2 Dallas TX 75265 @3 USA @Z 2 aut.`
A18	`01`	`1`		`@1 IEEE @2 New York NY @3 USA @9 patr.`
A20				`@1 839-842`
A21				`@1 1997`
A23	`01`			`@0 ENG`
A25	`01`			`@1 IEEE Computer Society Press @2 Washington DC`
A26	`01`			`@0 0-8186-7919-0`
A30	`01`	`1`	`ENG`	`@1 International conference on acoustics, speech, and signal processing @3 Munich DEU @4 1997-04-21`
A43	`01`			`@1 INIST @2 Y 31703 @5 354000077523980330`
A44				`@0 0000 @1 © 1998 INIST-CNRS. All rights reserved.`
A45				`@0 9 ref.`
A47	`01`	`1`		`@0 98-0082849`
A60				`@1 C`
A61				`@0 A`
A66	`01`			`@0 USA`
C01	`01`		`ENG`	@0 In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.
C02	`01`	`X`		`@0 001B40C60`
C02	`02`	`X`		`@0 001D04A05B`
C03	`01`	`3`	`FRE`	`@0 Traitement signal @5 01`
C03	`01`	`3`	`ENG`	`@0 Signal processing @5 01`
C03	`02`	`3`	`FRE`	`@0 Rapport signal bruit @5 02`
C03	`02`	`3`	`ENG`	`@0 Signal-to-noise ratio @5 02`
C03	`03`	`3`	`FRE`	`@0 Processus Markov @5 03`
C03	`03`	`3`	`ENG`	`@0 Markov process @5 03`
C03	`04`	`3`	`FRE`	`@0 Processus gaussien @5 04`
C03	`04`	`3`	`ENG`	`@0 Gaussian processes @5 04`
C03	`05`	`3`	`FRE`	`@0 Reconnaissance parole @5 05`
C03	`05`	`3`	`ENG`	`@0 Speech recognition @5 05`
C03	`06`	`X`	`FRE`	`@0 Maximum vraisemblance @5 06`
C03	`06`	`X`	`ENG`	`@0 Maximum likelihood @5 06`
C03	`06`	`X`	`SPA`	`@0 Maxima verosimilitud @5 06`
C03	`07`	`X`	`FRE`	`@0 Erreur systématique @5 08`
C03	`07`	`X`	`ENG`	`@0 Bias @5 08`
C03	`07`	`X`	`GER`	`@0 Systematischer Fehler @5 08`
C03	`07`	`X`	`SPA`	`@0 Error sistemático @5 08`
C03	`08`	`3`	`FRE`	`@0 4360 @2 PAC @4 INC @5 56`
C03	`09`	`3`	`FRE`	`@0 4372 @2 PAC @4 INC @5 57`
N21				`@1 047`

Format Inist (serveur)

NO :	PASCAL 98-0082849 INIST
ET :	A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition
AU :	AFIFY (M.); GONG (Y.); HATON (J.-P.)
AF :	CRIN/CNRS-INRIA-Lorraine, B.P. 239/54506 Vandeouvre, Nancy/France (1 aut., 3 aut.); Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374/Dallas TX 75265/Etats-Unis (2 aut.)
DT :	Congrès; Niveau analytique
SO :	International conference on acoustics, speech, and signal processing/1997-04-21/Munich DEU; Etats-Unis; Washington DC: IEEE Computer Society Press; Da. 1997; Pp. 839-842; ISBN 0-8186-7919-0
LA :	Anglais
EA :	In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.
CC :	001B40C60; 001D04A05B
FD :	Traitement signal; Rapport signal bruit; Processus Markov; Processus gaussien; Reconnaissance parole; Maximum vraisemblance; Erreur systématique; 4360; 4372
ED :	Signal processing; Signal-to-noise ratio; Markov process; Gaussian processes; Speech recognition; Maximum likelihood; Bias
GD :	Systematischer Fehler
SD :	Maxima verosimilitud; Error sistemático
LO :	INIST-Y 31703.354000077523980330
ID :	98-0082849

Links to Exploration step

Pascal:98-0082849

Le document en format XML

<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition</title>
<author><name sortKey="Afify, M" sort="Afify, M" uniqKey="Afify M" first="M." last="Afify">M. Afify</name>
<affiliation><inist:fA14 i1="01"><s1>CRIN/CNRS-INRIA-Lorraine, B.P. 239</s1>
<s2>54506 Vandeouvre, Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Gong, Y" sort="Gong, Y" uniqKey="Gong Y" first="Y." last="Gong">Y. Gong</name>
<affiliation><inist:fA14 i1="02"><s1>Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374</s1>
<s2>Dallas TX 75265</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Haton, J P" sort="Haton, J P" uniqKey="Haton J" first="J.-P." last="Haton">J.-P. Haton</name>
<affiliation><inist:fA14 i1="01"><s1>CRIN/CNRS-INRIA-Lorraine, B.P. 239</s1>
<s2>54506 Vandeouvre, Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">98-0082849</idno>
<date when="1997">1997</date>
<idno type="stanalyst">PASCAL 98-0082849 INIST</idno>
<idno type="RBID">Pascal:98-0082849</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000C14</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition</title>
<author><name sortKey="Afify, M" sort="Afify, M" uniqKey="Afify M" first="M." last="Afify">M. Afify</name>
<affiliation><inist:fA14 i1="01"><s1>CRIN/CNRS-INRIA-Lorraine, B.P. 239</s1>
<s2>54506 Vandeouvre, Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Gong, Y" sort="Gong, Y" uniqKey="Gong Y" first="Y." last="Gong">Y. Gong</name>
<affiliation><inist:fA14 i1="02"><s1>Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374</s1>
<s2>Dallas TX 75265</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Haton, J P" sort="Haton, J P" uniqKey="Haton J" first="J.-P." last="Haton">J.-P. Haton</name>
<affiliation><inist:fA14 i1="01"><s1>CRIN/CNRS-INRIA-Lorraine, B.P. 239</s1>
<s2>54506 Vandeouvre, Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
</analytic>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Bias</term>
<term>Gaussian processes</term>
<term>Markov process</term>
<term>Maximum likelihood</term>
<term>Signal processing</term>
<term>Signal-to-noise ratio</term>
<term>Speech recognition</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr"><term>Traitement signal</term>
<term>Rapport signal bruit</term>
<term>Processus Markov</term>
<term>Processus gaussien</term>
<term>Reconnaissance parole</term>
<term>Maximum vraisemblance</term>
<term>Erreur systématique</term>
<term>4360</term>
<term>4372</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.</div>
</front>
</TEI>
<inist><standard h6="B"><pA><fA08 i1="01" i2="1" l="ENG"><s1>A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition</s1>
</fA08>
<fA09 i1="01" i2="1" l="ENG"><s1>ICASSP 97 : international conference on acoustics, speech, and signal processing : Munich, April 21-24, 1997. Volume II: Speech processing</s1>
</fA09>
<fA11 i1="01" i2="1"><s1>AFIFY (M.)</s1>
</fA11>
<fA11 i1="02" i2="1"><s1>GONG (Y.)</s1>
</fA11>
<fA11 i1="03" i2="1"><s1>HATON (J.-P.)</s1>
</fA11>
<fA14 i1="01"><s1>CRIN/CNRS-INRIA-Lorraine, B.P. 239</s1>
<s2>54506 Vandeouvre, Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>3 aut.</sZ>
</fA14>
<fA14 i1="02"><s1>Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374</s1>
<s2>Dallas TX 75265</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
</fA14>
<fA18 i1="01" i2="1"><s1>IEEE</s1>
<s2>New York NY</s2>
<s3>USA</s3>
<s9>patr.</s9>
</fA18>
<fA20><s1>839-842</s1>
</fA20>
<fA21><s1>1997</s1>
</fA21>
<fA23 i1="01"><s0>ENG</s0>
</fA23>
<fA25 i1="01"><s1>IEEE Computer Society Press</s1>
<s2>Washington DC</s2>
</fA25>
<fA26 i1="01"><s0>0-8186-7919-0</s0>
</fA26>
<fA30 i1="01" i2="1" l="ENG"><s1>International conference on acoustics, speech, and signal processing</s1>
<s3>Munich DEU</s3>
<s4>1997-04-21</s4>
</fA30>
<fA43 i1="01"><s1>INIST</s1>
<s2>Y 31703</s2>
<s5>354000077523980330</s5>
</fA43>
<fA44><s0>0000</s0>
<s1>© 1998 INIST-CNRS. All rights reserved.</s1>
</fA44>
<fA45><s0>9 ref.</s0>
</fA45>
<fA47 i1="01" i2="1"><s0>98-0082849</s0>
</fA47>
<fA60><s1>C</s1>
</fA60>
<fA61><s0>A</s0>
</fA61>
<fA66 i1="01"><s0>USA</s0>
</fA66>
<fC01 i1="01" l="ENG"><s0>In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.</s0>
</fC01>
<fC02 i1="01" i2="X"><s0>001B40C60</s0>
</fC02>
<fC02 i1="02" i2="X"><s0>001D04A05B</s0>
</fC02>
<fC03 i1="01" i2="3" l="FRE"><s0>Traitement signal</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="3" l="ENG"><s0>Signal processing</s0>
<s5>01</s5>
</fC03>
<fC03 i1="02" i2="3" l="FRE"><s0>Rapport signal bruit</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="3" l="ENG"><s0>Signal-to-noise ratio</s0>
<s5>02</s5>
</fC03>
<fC03 i1="03" i2="3" l="FRE"><s0>Processus Markov</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="3" l="ENG"><s0>Markov process</s0>
<s5>03</s5>
</fC03>
<fC03 i1="04" i2="3" l="FRE"><s0>Processus gaussien</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="3" l="ENG"><s0>Gaussian processes</s0>
<s5>04</s5>
</fC03>
<fC03 i1="05" i2="3" l="FRE"><s0>Reconnaissance parole</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="3" l="ENG"><s0>Speech recognition</s0>
<s5>05</s5>
</fC03>
<fC03 i1="06" i2="X" l="FRE"><s0>Maximum vraisemblance</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="ENG"><s0>Maximum likelihood</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="SPA"><s0>Maxima verosimilitud</s0>
<s5>06</s5>
</fC03>
<fC03 i1="07" i2="X" l="FRE"><s0>Erreur systématique</s0>
<s5>08</s5>
</fC03>
<fC03 i1="07" i2="X" l="ENG"><s0>Bias</s0>
<s5>08</s5>
</fC03>
<fC03 i1="07" i2="X" l="GER"><s0>Systematischer Fehler</s0>
<s5>08</s5>
</fC03>
<fC03 i1="07" i2="X" l="SPA"><s0>Error sistemático</s0>
<s5>08</s5>
</fC03>
<fC03 i1="08" i2="3" l="FRE"><s0>4360</s0>
<s2>PAC</s2>
<s4>INC</s4>
<s5>56</s5>
</fC03>
<fC03 i1="09" i2="3" l="FRE"><s0>4372</s0>
<s2>PAC</s2>
<s4>INC</s4>
<s5>57</s5>
</fC03>
<fN21><s1>047</s1>
</fN21>
</pA>
</standard>
<server><NO>PASCAL 98-0082849 INIST</NO>
<ET>A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition</ET>
<AU>AFIFY (M.); GONG (Y.); HATON (J.-P.)</AU>
<AF>CRIN/CNRS-INRIA-Lorraine, B.P. 239/54506 Vandeouvre, Nancy/France (1 aut., 3 aut.); Speech Research, Personal Systems Laboratory, Texas Instruments, P.O. BOX 655303 MS 8374/Dallas TX 75265/Etats-Unis (2 aut.)</AF>
<DT>Congrès; Niveau analytique</DT>
<SO>International conference on acoustics, speech, and signal processing/1997-04-21/Munich DEU; Etats-Unis; Washington DC: IEEE Computer Society Press; Da. 1997; Pp. 839-842; ISBN 0-8186-7919-0</SO>
<LA>Anglais</LA>
<EA>In the context of continuous density hidden Markov model (CDHMM) we present a unified maximum likelihood (ML) approach to acoustic mismatch compensation. This is achieved by introducing additive Gaussian biases at the state level in both the mel cepstral and linear spectral domains. Flexible modelling of different mismatch effects can be obtained through appropriate bias tying. A Maximum likelihood approach for joint estimation of both mel cepstral and linear spectral biases from the observed mismatched speech given only one set of clean speech models is presented, where the obtained bias estimates are used for the compensation of clean speech models during decoding. The proposed approach is applied to the recognition of noisy Lombard speech, and significant improvement in the word recognition rate is achieved.</EA>
<CC>001B40C60; 001D04A05B</CC>
<FD>Traitement signal; Rapport signal bruit; Processus Markov; Processus gaussien; Reconnaissance parole; Maximum vraisemblance; Erreur systématique; 4360; 4372</FD>
<ED>Signal processing; Signal-to-noise ratio; Markov process; Gaussian processes; Speech recognition; Maximum likelihood; Bias</ED>
<GD>Systematischer Fehler</GD>
<SD>Maxima verosimilitud; Error sistemático</SD>
<LO>INIST-Y 31703.354000077523980330</LO>
<ID>98-0082849</ID>
</server>
</inist>
</record>

Pour manipuler ce document sous Unix (Dilib)

EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/PascalFrancis/Corpus

HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000C14 | SxmlIndent | more

HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Corpus/biblio.hfd -nk 000C14 | SxmlIndent | more

Pour mettre un lien sur cette page dans le réseau Wicri

{{Explor lien
   |wiki=    Wicri/Lorraine
   |area=    InforLorV4
   |flux=    PascalFrancis
   |étape=   Corpus
   |type=    RBID
   |clé=     Pascal:98-0082849
   |texte=   A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition
}}

This area was generated with Dilib version V0.6.33.
Data generation: Mon Jun 10 21:56:28 2019. Site generation: Fri Feb 25 15:29:27 2022

	Serveur d'exploration sur la recherche en informatique en Lorraine
	Attention, ce site est en cours de développement ! Attention, site généré par des moyens informatiques à partir de corpus bruts. Les informations ne sont donc pas validées.

Serveur d'exploration sur la recherche en informatique en Lorraine

A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition

A unified maximum likelihood approach to acoustic mismatch compensation : Application to noisy lombard speech recognition

Source :

Descripteurs français

English descriptors

Abstract

Notice en format standard (ISO 2709)

Format Inist (serveur)

Links to Exploration step

Le document en format XML

Pour manipuler ce document sous Unix (Dilib)

Pour mettre un lien sur cette page dans le réseau Wicri