α-Jacobian environmental adaptation
Identifieur interne : 000364 ( PascalFrancis/Curation ); précédent : 000363; suivant : 000365α-Jacobian environmental adaptation
Auteurs : Christophe Cerisara [France] ; Luca Rigazio [États-Unis] ; Jean-Claude Junqua [États-Unis]Source :
- Speech communication [ 0167-6393 ] ; 2004.
Descripteurs français
- Pascal (Inist)
English descriptors
- KwdEn :
Abstract
The robustness of automatic speech recognition systems to noise is still a problem, especially for small footprint systems. This paper addresses the problem of noise robustness using model compensation methods. Such algorithms are already available, but their complexity is usually high. An often-referenced method for achieving noise robustness is parallel model combination (PMC). Several algorithms have been proposed to develop more computationally efficient methods than PMC. For example, Jacobian adaptation approximates PMC with a linear transformation function in the cepstral domain. However, the Jacobian approximation is valid only for test environments that are close to the training conditions whereas, in real test conditions, the mismatch between the test and training environments is usually large. In this paper, we propose two methods, respectively called static and dynamic α-Jacobian adaptation (or α-JAC), to compute new linear approximations of PMC for realistic test environments. We further extend both algorithms to compensate for additive and convolutional noise and we derive the corresponding non-linear algorithm that is approximated. All these algorithms are experimentally compared in important mismatch conditions. As compared to Jacobian adaptation, improvements are observed with both static and dynamic α-Jacobian adaptation.
pA |
|
---|
Links toward previous steps (curation, corpus...)
- to stream PascalFrancis, to step Corpus: Pour aller vers cette notice dans l'étape Curation :000677
Links to Exploration step
Pascal:04-0274368Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">α-Jacobian environmental adaptation</title>
<author><name sortKey="Cerisara, Christophe" sort="Cerisara, Christophe" uniqKey="Cerisara C" first="Christophe" last="Cerisara">Christophe Cerisara</name>
<affiliation wicri:level="1"><inist:fA14 i1="01"><s1>LORIA, UMR 7503 Campus Scientifique, BP 239</s1>
<s2>54506 Vandoeuvre-les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14>
<country>France</country>
</affiliation>
</author>
<author><name sortKey="Rigazio, Luca" sort="Rigazio, Luca" uniqKey="Rigazio L" first="Luca" last="Rigazio">Luca Rigazio</name>
<affiliation wicri:level="1"><inist:fA14 i1="02"><s1>Panasonic Speech Technology Laboratory, 3888 State Street, Suite 202</s1>
<s2>Santa-Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>États-Unis</country>
</affiliation>
</author>
<author><name sortKey="Junqua, Jean Claude" sort="Junqua, Jean Claude" uniqKey="Junqua J" first="Jean-Claude" last="Junqua">Jean-Claude Junqua</name>
<affiliation wicri:level="1"><inist:fA14 i1="02"><s1>Panasonic Speech Technology Laboratory, 3888 State Street, Suite 202</s1>
<s2>Santa-Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>États-Unis</country>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">04-0274368</idno>
<date when="2004">2004</date>
<idno type="stanalyst">PASCAL 04-0274368 INIST</idno>
<idno type="RBID">Pascal:04-0274368</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000677</idno>
<idno type="wicri:Area/PascalFrancis/Curation">000364</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">α-Jacobian environmental adaptation</title>
<author><name sortKey="Cerisara, Christophe" sort="Cerisara, Christophe" uniqKey="Cerisara C" first="Christophe" last="Cerisara">Christophe Cerisara</name>
<affiliation wicri:level="1"><inist:fA14 i1="01"><s1>LORIA, UMR 7503 Campus Scientifique, BP 239</s1>
<s2>54506 Vandoeuvre-les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</inist:fA14>
<country>France</country>
</affiliation>
</author>
<author><name sortKey="Rigazio, Luca" sort="Rigazio, Luca" uniqKey="Rigazio L" first="Luca" last="Rigazio">Luca Rigazio</name>
<affiliation wicri:level="1"><inist:fA14 i1="02"><s1>Panasonic Speech Technology Laboratory, 3888 State Street, Suite 202</s1>
<s2>Santa-Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>États-Unis</country>
</affiliation>
</author>
<author><name sortKey="Junqua, Jean Claude" sort="Junqua, Jean Claude" uniqKey="Junqua J" first="Jean-Claude" last="Junqua">Jean-Claude Junqua</name>
<affiliation wicri:level="1"><inist:fA14 i1="02"><s1>Panasonic Speech Technology Laboratory, 3888 State Street, Suite 202</s1>
<s2>Santa-Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
<sZ>3 aut.</sZ>
</inist:fA14>
<country>États-Unis</country>
</affiliation>
</author>
</analytic>
<series><title level="j" type="main">Speech communication</title>
<title level="j" type="abbreviated">Speech commun.</title>
<idno type="ISSN">0167-6393</idno>
<imprint><date when="2004">2004</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
<seriesStmt><title level="j" type="main">Speech communication</title>
<title level="j" type="abbreviated">Speech commun.</title>
<idno type="ISSN">0167-6393</idno>
</seriesStmt>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Adaptive method</term>
<term>Additive noise</term>
<term>Automatic recognition</term>
<term>Jacobi method</term>
<term>Noise reduction</term>
<term>Robustness</term>
<term>Speech recognition</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr"><term>Reconnaissance parole</term>
<term>Reconnaissance automatique</term>
<term>Robustesse</term>
<term>Méthode Jacobi</term>
<term>Méthode adaptative</term>
<term>Réduction bruit</term>
<term>Bruit additif</term>
<term>Compensation modèle</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">The robustness of automatic speech recognition systems to noise is still a problem, especially for small footprint systems. This paper addresses the problem of noise robustness using model compensation methods. Such algorithms are already available, but their complexity is usually high. An often-referenced method for achieving noise robustness is parallel model combination (PMC). Several algorithms have been proposed to develop more computationally efficient methods than PMC. For example, Jacobian adaptation approximates PMC with a linear transformation function in the cepstral domain. However, the Jacobian approximation is valid only for test environments that are close to the training conditions whereas, in real test conditions, the mismatch between the test and training environments is usually large. In this paper, we propose two methods, respectively called static and dynamic α-Jacobian adaptation (or α-JAC), to compute new linear approximations of PMC for realistic test environments. We further extend both algorithms to compensate for additive and convolutional noise and we derive the corresponding non-linear algorithm that is approximated. All these algorithms are experimentally compared in important mismatch conditions. As compared to Jacobian adaptation, improvements are observed with both static and dynamic α-Jacobian adaptation.</div>
</front>
</TEI>
<inist><standard h6="B"><pA><fA01 i1="01" i2="1"><s0>0167-6393</s0>
</fA01>
<fA02 i1="01"><s0>SCOMDH</s0>
</fA02>
<fA03 i2="1"><s0>Speech commun.</s0>
</fA03>
<fA05><s2>42</s2>
</fA05>
<fA06><s2>1</s2>
</fA06>
<fA08 i1="01" i2="1" l="ENG"><s1>α-Jacobian environmental adaptation</s1>
</fA08>
<fA09 i1="01" i2="1" l="ENG"><s1>Adaptation methods for speech recognition</s1>
</fA09>
<fA11 i1="01" i2="1"><s1>CERISARA (Christophe)</s1>
</fA11>
<fA11 i1="02" i2="1"><s1>RIGAZIO (Luca)</s1>
</fA11>
<fA11 i1="03" i2="1"><s1>JUNQUA (Jean-Claude)</s1>
</fA11>
<fA12 i1="01" i2="1"><s1>JUNQUA (J. C.)</s1>
<s9>ed.</s9>
</fA12>
<fA12 i1="02" i2="1"><s1>WELLEKENS (C. J.)</s1>
<s9>ed.</s9>
</fA12>
<fA14 i1="01"><s1>LORIA, UMR 7503 Campus Scientifique, BP 239</s1>
<s2>54506 Vandoeuvre-les-Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
</fA14>
<fA14 i1="02"><s1>Panasonic Speech Technology Laboratory, 3888 State Street, Suite 202</s1>
<s2>Santa-Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>2 aut.</sZ>
<sZ>3 aut.</sZ>
</fA14>
<fA15 i1="01"><s1>Panasonic Speech Technology Laboratory of Panasonic Tech. Company, Div. of Matsushita Electric Corp. of America, Suite 202, 3888 State Street</s1>
<s2>Santa Barbara, CA 93105</s2>
<s3>USA</s3>
<sZ>1 aut.</sZ>
</fA15>
<fA15 i1="02"><s1>Institut Eurécom, Multimedia Communications Dpt, 2229, Route de Cretes - B.P. 193</s1>
<s2>06904 Sophia Antipolis</s2>
<s3>FRA</s3>
<sZ>2 aut.</sZ>
</fA15>
<fA20><s1>25-41</s1>
</fA20>
<fA21><s1>2004</s1>
</fA21>
<fA23 i1="01"><s0>ENG</s0>
</fA23>
<fA43 i1="01"><s1>INIST</s1>
<s2>19642</s2>
<s5>354000116515290020</s5>
</fA43>
<fA44><s0>0000</s0>
<s1>© 2004 INIST-CNRS. All rights reserved.</s1>
</fA44>
<fA45><s0>1 p.1/4</s0>
</fA45>
<fA47 i1="01" i2="1"><s0>04-0274368</s0>
</fA47>
<fA60><s1>P</s1>
</fA60>
<fA61><s0>A</s0>
</fA61>
<fA64 i1="01" i2="1"><s0>Speech communication</s0>
</fA64>
<fA66 i1="01"><s0>NLD</s0>
</fA66>
<fC01 i1="01" l="ENG"><s0>The robustness of automatic speech recognition systems to noise is still a problem, especially for small footprint systems. This paper addresses the problem of noise robustness using model compensation methods. Such algorithms are already available, but their complexity is usually high. An often-referenced method for achieving noise robustness is parallel model combination (PMC). Several algorithms have been proposed to develop more computationally efficient methods than PMC. For example, Jacobian adaptation approximates PMC with a linear transformation function in the cepstral domain. However, the Jacobian approximation is valid only for test environments that are close to the training conditions whereas, in real test conditions, the mismatch between the test and training environments is usually large. In this paper, we propose two methods, respectively called static and dynamic α-Jacobian adaptation (or α-JAC), to compute new linear approximations of PMC for realistic test environments. We further extend both algorithms to compensate for additive and convolutional noise and we derive the corresponding non-linear algorithm that is approximated. All these algorithms are experimentally compared in important mismatch conditions. As compared to Jacobian adaptation, improvements are observed with both static and dynamic α-Jacobian adaptation.</s0>
</fC01>
<fC02 i1="01" i2="X"><s0>001D04A05B</s0>
</fC02>
<fC03 i1="01" i2="X" l="FRE"><s0>Reconnaissance parole</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="ENG"><s0>Speech recognition</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="SPA"><s0>Reconocimiento voz</s0>
<s5>01</s5>
</fC03>
<fC03 i1="02" i2="X" l="FRE"><s0>Reconnaissance automatique</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="ENG"><s0>Automatic recognition</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="SPA"><s0>Reconocimiento automático</s0>
<s5>02</s5>
</fC03>
<fC03 i1="03" i2="X" l="FRE"><s0>Robustesse</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="ENG"><s0>Robustness</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="SPA"><s0>Robustez</s0>
<s5>03</s5>
</fC03>
<fC03 i1="04" i2="X" l="FRE"><s0>Méthode Jacobi</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="ENG"><s0>Jacobi method</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="SPA"><s0>Método Jacobi</s0>
<s5>04</s5>
</fC03>
<fC03 i1="05" i2="X" l="FRE"><s0>Méthode adaptative</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="ENG"><s0>Adaptive method</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="SPA"><s0>Método adaptativo</s0>
<s5>05</s5>
</fC03>
<fC03 i1="06" i2="X" l="FRE"><s0>Réduction bruit</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="ENG"><s0>Noise reduction</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="SPA"><s0>Reducción ruido</s0>
<s5>06</s5>
</fC03>
<fC03 i1="07" i2="X" l="FRE"><s0>Bruit additif</s0>
<s5>07</s5>
</fC03>
<fC03 i1="07" i2="X" l="ENG"><s0>Additive noise</s0>
<s5>07</s5>
</fC03>
<fC03 i1="07" i2="X" l="SPA"><s0>Ruido aditivo</s0>
<s5>07</s5>
</fC03>
<fC03 i1="08" i2="X" l="FRE"><s0>Compensation modèle</s0>
<s4>INC</s4>
<s5>72</s5>
</fC03>
<fN21><s1>173</s1>
</fN21>
<fN44 i1="01"><s1>PSI</s1>
</fN44>
<fN82><s1>PSI</s1>
</fN82>
</pA>
</standard>
</inist>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/PascalFrancis/Curation
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000364 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Curation/biblio.hfd -nk 000364 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Lorraine |area= InforLorV4 |flux= PascalFrancis |étape= Curation |type= RBID |clé= Pascal:04-0274368 |texte= α-Jacobian environmental adaptation }}
This area was generated with Dilib version V0.6.33. |