InforLorV4, PascalFrancis, Corpus, bibRecord, 000428

Optimizing the coverage of a speech database through a selection of representative speaker recordings

Identifieur interne : 000428 ( PascalFrancis/Corpus ); précédent : 000427; suivant : 000429

Optimizing the coverage of a speech database through a selection of representative speaker recordings

Auteurs : Sacha Krstulovic ; Frédéric Bimbot ; Olivier Boëffard ; Delphine Charlet ; Dominique Fohr ; Odile Mella

Source :

Speech communication [ 0167-6393 ] ; 2006.

RBID : Pascal:06-0450665

Descripteurs français

Pascal (Inist)
- Optimisation, Base donnée, Français, Critère qualité, Similitude, Modélisation, Classification automatique, Algorithme, Classification signal.

English descriptors

KwdEn :
- Algorithm, Automatic classification, Database, French, Modeling, Optimization, Quality criterion, Signal classification, Similarity.

Abstract

In the context of the NEOLOGOS French speech database creation project,¹ a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.

Notice en format standard (ISO 2709)

Pour connaître la documentation sur le format Inist Standard.

A01	`01`	`1`		`@0 0167-6393`
A02	`01`			`@0 SCOMDH`
A03		`1`		`@0 Speech commun.`
A05				`@2 48`
A06				`@2 10`
A08	`01`	`1`	`ENG`	`@1 Optimizing the coverage of a speech database through a selection of representative speaker recordings`
A11	`01`	`1`		`@1 KRSTULOVIC (Sacha)`
A11	`02`	`1`		`@1 BIMBOT (Frédéric)`
A11	`03`	`1`		`@1 BOËFFARD (Olivier)`
A11	`04`	`1`		`@1 CHARLET (Delphine)`
A11	`05`	`1`		`@1 FOHR (Dominique)`
A11	`06`	`1`		`@1 MELLA (Odile)`
A14	`01`			`@1 IRISA/METISS, Campus de Beaulieu @2 35042 Rennes @3 FRA @Z 1 aut. @Z 2 aut.`
A14	`02`			`@1 IRISAICORDIAL, 6 r. Kerampont, BP 80518 @2 22305 Lannion @3 FRA @Z 3 aut.`
A14	`03`			`@1 France Télécom R&D, 2 ave. Marzin @2 22307 Lannion @3 FRA @Z 4 aut.`
A14	`04`			`@1 LORIA, Campus Universitaire, BP 239 @2 54506 Vandoeuvre @3 FRA @Z 5 aut. @Z 6 aut.`
A20				`@1 1319-1348`
A21				`@1 2006`
A23	`01`			`@0 ENG`
A43	`01`			`@1 INIST @2 19642 @5 354000158725920080`
A44				`@0 0000 @1 © 2006 INIST-CNRS. All rights reserved.`
A45				`@0 31 ref.`
A47	`01`	`1`		`@0 06-0450665`
A60				`@1 P`
A61				`@0 A`
A64	`01`	`1`		`@0 Speech communication`
A66	`01`			`@0 NLD`
C01	`01`		`ENG`	@0 In the context of the NEOLOGOS French speech database creation project,¹ a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.
C02	`01`	`X`		`@0 001D04A05B`
C02	`02`	`X`		`@0 001D04A04A1`
C03	`01`	`X`	`FRE`	`@0 Optimisation @5 01`
C03	`01`	`X`	`ENG`	`@0 Optimization @5 01`
C03	`01`	`X`	`SPA`	`@0 Optimización @5 01`
C03	`02`	`X`	`FRE`	`@0 Base donnée @5 02`
C03	`02`	`X`	`ENG`	`@0 Database @5 02`
C03	`02`	`X`	`SPA`	`@0 Base dato @5 02`
C03	`03`	`X`	`FRE`	`@0 Français @5 03`
C03	`03`	`X`	`ENG`	`@0 French @5 03`
C03	`03`	`X`	`SPA`	`@0 Francés @5 03`
C03	`04`	`X`	`FRE`	`@0 Critère qualité @5 04`
C03	`04`	`X`	`ENG`	`@0 Quality criterion @5 04`
C03	`04`	`X`	`SPA`	`@0 Criterio calidad @5 04`
C03	`05`	`X`	`FRE`	`@0 Similitude @5 05`
C03	`05`	`X`	`ENG`	`@0 Similarity @5 05`
C03	`05`	`X`	`SPA`	`@0 Similitud @5 05`
C03	`06`	`X`	`FRE`	`@0 Modélisation @5 06`
C03	`06`	`X`	`ENG`	`@0 Modeling @5 06`
C03	`06`	`X`	`SPA`	`@0 Modelización @5 06`
C03	`07`	`X`	`FRE`	`@0 Classification automatique @5 07`
C03	`07`	`X`	`ENG`	`@0 Automatic classification @5 07`
C03	`07`	`X`	`SPA`	`@0 Clasificación automática @5 07`
C03	`08`	`X`	`FRE`	`@0 Algorithme @5 08`
C03	`08`	`X`	`ENG`	`@0 Algorithm @5 08`
C03	`08`	`X`	`SPA`	`@0 Algoritmo @5 08`
C03	`09`	`3`	`FRE`	`@0 Classification signal @5 31`
C03	`09`	`3`	`ENG`	`@0 Signal classification @5 31`
N21				`@1 296`
N44	`01`			`@1 OTO`
N82				`@1 OTO`

Format Inist (serveur)

NO :	PASCAL 06-0450665 INIST
ET :	Optimizing the coverage of a speech database through a selection of representative speaker recordings
AU :	KRSTULOVIC (Sacha); BIMBOT (Frédéric); BOËFFARD (Olivier); CHARLET (Delphine); FOHR (Dominique); MELLA (Odile)
AF :	IRISA/METISS, Campus de Beaulieu/35042 Rennes/France (1 aut., 2 aut.); IRISAICORDIAL, 6 r. Kerampont, BP 80518/22305 Lannion/France (3 aut.); France Télécom R&D, 2 ave. Marzin/22307 Lannion/France (4 aut.); LORIA, Campus Universitaire, BP 239/54506 Vandoeuvre/France (5 aut., 6 aut.)
DT :	Publication en série; Niveau analytique
SO :	Speech communication; ISSN 0167-6393; Coden SCOMDH; Pays-Bas; Da. 2006; Vol. 48; No. 10; Pp. 1319-1348; Bibl. 31 ref.
LA :	Anglais
EA :	In the context of the NEOLOGOS French speech database creation project,¹ a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.
CC :	001D04A05B; 001D04A04A1
FD :	Optimisation; Base donnée; Français; Critère qualité; Similitude; Modélisation; Classification automatique; Algorithme; Classification signal
ED :	Optimization; Database; French; Quality criterion; Similarity; Modeling; Automatic classification; Algorithm; Signal classification
SD :	Optimización; Base dato; Francés; Criterio calidad; Similitud; Modelización; Clasificación automática; Algoritmo
LO :	INIST-19642.354000158725920080
ID :	06-0450665

Links to Exploration step

Pascal:06-0450665

Le document en format XML

<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">Optimizing the coverage of a speech database through a selection of representative speaker recordings</title>
<author><name sortKey="Krstulovic, Sacha" sort="Krstulovic, Sacha" uniqKey="Krstulovic S" first="Sacha" last="Krstulovic">Sacha Krstulovic</name>
<affiliation><inist:fA14 i1="01"><s1>IRISA/METISS, Campus de Beaulieu</s1>
<s2>35042 Rennes</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Bimbot, Frederic" sort="Bimbot, Frederic" uniqKey="Bimbot F" first="Frédéric" last="Bimbot">Frédéric Bimbot</name>
<affiliation><inist:fA14 i1="01"><s1>IRISA/METISS, Campus de Beaulieu</s1>
<s2>35042 Rennes</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Boeffard, Olivier" sort="Boeffard, Olivier" uniqKey="Boeffard O" first="Olivier" last="Boëffard">Olivier Boëffard</name>
<affiliation><inist:fA14 i1="02"><s1>IRISAICORDIAL, 6 r. Kerampont, BP 80518</s1>
<s2>22305 Lannion</s2>
<s3>FRA</s3>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Charlet, Delphine" sort="Charlet, Delphine" uniqKey="Charlet D" first="Delphine" last="Charlet">Delphine Charlet</name>
<affiliation><inist:fA14 i1="03"><s1>France Télécom R&D, 2 ave. Marzin</s1>
<s2>22307 Lannion</s2>
<s3>FRA</s3>
<sZ>4 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Fohr, Dominique" sort="Fohr, Dominique" uniqKey="Fohr D" first="Dominique" last="Fohr">Dominique Fohr</name>
<affiliation><inist:fA14 i1="04"><s1>LORIA, Campus Universitaire, BP 239</s1>
<s2>54506 Vandoeuvre</s2>
<s3>FRA</s3>
<sZ>5 aut.</sZ>
<sZ>6 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Mella, Odile" sort="Mella, Odile" uniqKey="Mella O" first="Odile" last="Mella">Odile Mella</name>
<affiliation><inist:fA14 i1="04"><s1>LORIA, Campus Universitaire, BP 239</s1>
<s2>54506 Vandoeuvre</s2>
<s3>FRA</s3>
<sZ>5 aut.</sZ>
<sZ>6 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">06-0450665</idno>
<date when="2006">2006</date>
<idno type="stanalyst">PASCAL 06-0450665 INIST</idno>
<idno type="RBID">Pascal:06-0450665</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000428</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">Optimizing the coverage of a speech database through a selection of representative speaker recordings</title>
<author><name sortKey="Krstulovic, Sacha" sort="Krstulovic, Sacha" uniqKey="Krstulovic S" first="Sacha" last="Krstulovic">Sacha Krstulovic</name>
<affiliation><inist:fA14 i1="01"><s1>IRISA/METISS, Campus de Beaulieu</s1>
<s2>35042 Rennes</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Bimbot, Frederic" sort="Bimbot, Frederic" uniqKey="Bimbot F" first="Frédéric" last="Bimbot">Frédéric Bimbot</name>
<affiliation><inist:fA14 i1="01"><s1>IRISA/METISS, Campus de Beaulieu</s1>
<s2>35042 Rennes</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Boeffard, Olivier" sort="Boeffard, Olivier" uniqKey="Boeffard O" first="Olivier" last="Boëffard">Olivier Boëffard</name>
<affiliation><inist:fA14 i1="02"><s1>IRISAICORDIAL, 6 r. Kerampont, BP 80518</s1>
<s2>22305 Lannion</s2>
<s3>FRA</s3>
<sZ>3 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Charlet, Delphine" sort="Charlet, Delphine" uniqKey="Charlet D" first="Delphine" last="Charlet">Delphine Charlet</name>
<affiliation><inist:fA14 i1="03"><s1>France Télécom R&D, 2 ave. Marzin</s1>
<s2>22307 Lannion</s2>
<s3>FRA</s3>
<sZ>4 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Fohr, Dominique" sort="Fohr, Dominique" uniqKey="Fohr D" first="Dominique" last="Fohr">Dominique Fohr</name>
<affiliation><inist:fA14 i1="04"><s1>LORIA, Campus Universitaire, BP 239</s1>
<s2>54506 Vandoeuvre</s2>
<s3>FRA</s3>
<sZ>5 aut.</sZ>
<sZ>6 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Mella, Odile" sort="Mella, Odile" uniqKey="Mella O" first="Odile" last="Mella">Odile Mella</name>
<affiliation><inist:fA14 i1="04"><s1>LORIA, Campus Universitaire, BP 239</s1>
<s2>54506 Vandoeuvre</s2>
<s3>FRA</s3>
<sZ>5 aut.</sZ>
<sZ>6 aut.</sZ>
</inist:fA14>
</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="2006">2006</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>Algorithm</term>
<term>Automatic classification</term>
<term>Database</term>
<term>French</term>
<term>Modeling</term>
<term>Optimization</term>
<term>Quality criterion</term>
<term>Signal classification</term>
<term>Similarity</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr"><term>Optimisation</term>
<term>Base donnée</term>
<term>Français</term>
<term>Critère qualité</term>
<term>Similitude</term>
<term>Modélisation</term>
<term>Classification automatique</term>
<term>Algorithme</term>
<term>Classification signal</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">In the context of the NEOLOGOS French speech database creation project,<sup>1</sup>
 a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.</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>48</s2>
</fA05>
<fA06><s2>10</s2>
</fA06>
<fA08 i1="01" i2="1" l="ENG"><s1>Optimizing the coverage of a speech database through a selection of representative speaker recordings</s1>
</fA08>
<fA11 i1="01" i2="1"><s1>KRSTULOVIC (Sacha)</s1>
</fA11>
<fA11 i1="02" i2="1"><s1>BIMBOT (Frédéric)</s1>
</fA11>
<fA11 i1="03" i2="1"><s1>BOËFFARD (Olivier)</s1>
</fA11>
<fA11 i1="04" i2="1"><s1>CHARLET (Delphine)</s1>
</fA11>
<fA11 i1="05" i2="1"><s1>FOHR (Dominique)</s1>
</fA11>
<fA11 i1="06" i2="1"><s1>MELLA (Odile)</s1>
</fA11>
<fA14 i1="01"><s1>IRISA/METISS, Campus de Beaulieu</s1>
<s2>35042 Rennes</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</fA14>
<fA14 i1="02"><s1>IRISAICORDIAL, 6 r. Kerampont, BP 80518</s1>
<s2>22305 Lannion</s2>
<s3>FRA</s3>
<sZ>3 aut.</sZ>
</fA14>
<fA14 i1="03"><s1>France Télécom R&D, 2 ave. Marzin</s1>
<s2>22307 Lannion</s2>
<s3>FRA</s3>
<sZ>4 aut.</sZ>
</fA14>
<fA14 i1="04"><s1>LORIA, Campus Universitaire, BP 239</s1>
<s2>54506 Vandoeuvre</s2>
<s3>FRA</s3>
<sZ>5 aut.</sZ>
<sZ>6 aut.</sZ>
</fA14>
<fA20><s1>1319-1348</s1>
</fA20>
<fA21><s1>2006</s1>
</fA21>
<fA23 i1="01"><s0>ENG</s0>
</fA23>
<fA43 i1="01"><s1>INIST</s1>
<s2>19642</s2>
<s5>354000158725920080</s5>
</fA43>
<fA44><s0>0000</s0>
<s1>© 2006 INIST-CNRS. All rights reserved.</s1>
</fA44>
<fA45><s0>31 ref.</s0>
</fA45>
<fA47 i1="01" i2="1"><s0>06-0450665</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>In the context of the NEOLOGOS French speech database creation project,<sup>1</sup>
 a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.</s0>
</fC01>
<fC02 i1="01" i2="X"><s0>001D04A05B</s0>
</fC02>
<fC02 i1="02" i2="X"><s0>001D04A04A1</s0>
</fC02>
<fC03 i1="01" i2="X" l="FRE"><s0>Optimisation</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="ENG"><s0>Optimization</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="SPA"><s0>Optimización</s0>
<s5>01</s5>
</fC03>
<fC03 i1="02" i2="X" l="FRE"><s0>Base donnée</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="ENG"><s0>Database</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="SPA"><s0>Base dato</s0>
<s5>02</s5>
</fC03>
<fC03 i1="03" i2="X" l="FRE"><s0>Français</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="ENG"><s0>French</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="SPA"><s0>Francés</s0>
<s5>03</s5>
</fC03>
<fC03 i1="04" i2="X" l="FRE"><s0>Critère qualité</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="ENG"><s0>Quality criterion</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="SPA"><s0>Criterio calidad</s0>
<s5>04</s5>
</fC03>
<fC03 i1="05" i2="X" l="FRE"><s0>Similitude</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="ENG"><s0>Similarity</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="SPA"><s0>Similitud</s0>
<s5>05</s5>
</fC03>
<fC03 i1="06" i2="X" l="FRE"><s0>Modélisation</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="ENG"><s0>Modeling</s0>
<s5>06</s5>
</fC03>
<fC03 i1="06" i2="X" l="SPA"><s0>Modelización</s0>
<s5>06</s5>
</fC03>
<fC03 i1="07" i2="X" l="FRE"><s0>Classification automatique</s0>
<s5>07</s5>
</fC03>
<fC03 i1="07" i2="X" l="ENG"><s0>Automatic classification</s0>
<s5>07</s5>
</fC03>
<fC03 i1="07" i2="X" l="SPA"><s0>Clasificación automática</s0>
<s5>07</s5>
</fC03>
<fC03 i1="08" i2="X" l="FRE"><s0>Algorithme</s0>
<s5>08</s5>
</fC03>
<fC03 i1="08" i2="X" l="ENG"><s0>Algorithm</s0>
<s5>08</s5>
</fC03>
<fC03 i1="08" i2="X" l="SPA"><s0>Algoritmo</s0>
<s5>08</s5>
</fC03>
<fC03 i1="09" i2="3" l="FRE"><s0>Classification signal</s0>
<s5>31</s5>
</fC03>
<fC03 i1="09" i2="3" l="ENG"><s0>Signal classification</s0>
<s5>31</s5>
</fC03>
<fN21><s1>296</s1>
</fN21>
<fN44 i1="01"><s1>OTO</s1>
</fN44>
<fN82><s1>OTO</s1>
</fN82>
</pA>
</standard>
<server><NO>PASCAL 06-0450665 INIST</NO>
<ET>Optimizing the coverage of a speech database through a selection of representative speaker recordings</ET>
<AU>KRSTULOVIC (Sacha); BIMBOT (Frédéric); BOËFFARD (Olivier); CHARLET (Delphine); FOHR (Dominique); MELLA (Odile)</AU>
<AF>IRISA/METISS, Campus de Beaulieu/35042 Rennes/France (1 aut., 2 aut.); IRISAICORDIAL, 6 r. Kerampont, BP 80518/22305 Lannion/France (3 aut.); France Télécom R&D, 2 ave. Marzin/22307 Lannion/France (4 aut.); LORIA, Campus Universitaire, BP 239/54506 Vandoeuvre/France (5 aut., 6 aut.)</AF>
<DT>Publication en série; Niveau analytique</DT>
<SO>Speech communication; ISSN 0167-6393; Coden SCOMDH; Pays-Bas; Da. 2006; Vol. 48; No. 10; Pp. 1319-1348; Bibl. 31 ref.</SO>
<LA>Anglais</LA>
<EA>In the context of the NEOLOGOS French speech database creation project,<sup>1</sup>
 a general methodology was defined for the selection of representative speaker recordings. The selection aims at providing a good coverage in terms of speaker variability while limiting the number of recorded speakers. This is intended to make the resulting database both more adapted to the development of recently proposed multi-model methods and less expensive to collect. The presented methodology proposes a selection process based on the optimization of a quality criterion defined in a variety of speaker similarity modeling frameworks. The selection can be achieved with respect to a unique similarity criterion, using classical clustering methods such as hierarchical or K-medians clustering, or it can combine several speaker similarity criteria, thanks to a newly developed clustering method called focal speakers selection. In this framework, four different speaker similarity criteria are tested, and three different speaker clustering algorithms are compared. Results pertaining to the collection of the NEOLOGOS database are also discussed.</EA>
<CC>001D04A05B; 001D04A04A1</CC>
<FD>Optimisation; Base donnée; Français; Critère qualité; Similitude; Modélisation; Classification automatique; Algorithme; Classification signal</FD>
<ED>Optimization; Database; French; Quality criterion; Similarity; Modeling; Automatic classification; Algorithm; Signal classification</ED>
<SD>Optimización; Base dato; Francés; Criterio calidad; Similitud; Modelización; Clasificación automática; Algoritmo</SD>
<LO>INIST-19642.354000158725920080</LO>
<ID>06-0450665</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 000428 | SxmlIndent | more

HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Corpus/biblio.hfd -nk 000428 | 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:06-0450665
   |texte=   Optimizing the coverage of a speech database through a selection of representative speaker recordings
}}

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

Optimizing the coverage of a speech database through a selection of representative speaker recordings

Optimizing the coverage of a speech database through a selection of representative speaker recordings

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