Frequency and wavelet filtering for robust speech recognition
Identifieur interne :
000727 ( PascalFrancis/Corpus );
précédent :
000726;
suivant :
000728
Frequency and wavelet filtering for robust speech recognition
Auteurs : Murat Deviren ;
Khalid DaoudiSource :
-
Lecture notes in computer science [ 0302-9743 ] ; 2003.
RBID : Pascal:04-0130632
Descripteurs français
English descriptors
Abstract
Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case.
Notice en format standard (ISO 2709)
Pour connaître la documentation sur le format Inist Standard.
pA |
A01 | 01 | 1 | | @0 0302-9743 |
---|
A05 | | | | @2 2714 |
---|
A08 | 01 | 1 | ENG | @1 Frequency and wavelet filtering for robust speech recognition |
---|
A09 | 01 | 1 | ENG | @1 Artificial neural networks and neural information processing - ICANN / ICONIP 2003 : Istanbul, 26-29 June 2003 |
---|
A11 | 01 | 1 | | @1 DEVIREN (Murat) |
---|
A11 | 02 | 1 | | @1 DAOUDI (Khalid) |
---|
A12 | 01 | 1 | | @1 KAYNAK (Okyay) @9 ed. |
---|
A12 | 02 | 1 | | @1 ALPAYDIN (Ethem) @9 ed. |
---|
A12 | 03 | 1 | | @1 OJA (Erkki) @9 ed. |
---|
A12 | 04 | 1 | | @1 LEI XU @9 ed. |
---|
A14 | 01 | | | @1 INRIA-LORIA (Speech Group) B.P. 101 @2 54602 Villers les Nancy @3 FRA @Z 1 aut. @Z 2 aut. |
---|
A20 | | | | @1 452-460 |
---|
A21 | | | | @1 2003 |
---|
A23 | 01 | | | @0 ENG |
---|
A26 | 01 | | | @0 3-540-40408-2 |
---|
A43 | 01 | | | @1 INIST @2 16343 @5 354000117807160540 |
---|
A44 | | | | @0 0000 @1 © 2004 INIST-CNRS. All rights reserved. |
---|
A45 | | | | @0 11 ref. |
---|
A47 | 01 | 1 | | @0 04-0130632 |
---|
A60 | | | | @1 P @2 C |
---|
A61 | | | | @0 A |
---|
A64 | 01 | 1 | | @0 Lecture notes in computer science |
---|
A66 | 01 | | | @0 DEU |
---|
C01 | 01 | | ENG | @0 Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case. |
---|
C02 | 01 | X | | @0 001D02C04 |
---|
C02 | 02 | X | | @0 001D04A05B |
---|
C03 | 01 | X | FRE | @0 Intelligence artificielle @5 01 |
---|
C03 | 01 | X | ENG | @0 Artificial intelligence @5 01 |
---|
C03 | 01 | X | SPA | @0 Inteligencia artificial @5 01 |
---|
C03 | 02 | X | FRE | @0 Interface utilisateur @5 02 |
---|
C03 | 02 | X | ENG | @0 User interface @5 02 |
---|
C03 | 02 | X | SPA | @0 Interfase usuario @5 02 |
---|
C03 | 03 | X | FRE | @0 Base donnée @5 03 |
---|
C03 | 03 | X | ENG | @0 Database @5 03 |
---|
C03 | 03 | X | SPA | @0 Base dato @5 03 |
---|
C03 | 04 | X | FRE | @0 Reconnaissance parole @5 04 |
---|
C03 | 04 | X | ENG | @0 Speech recognition @5 04 |
---|
C03 | 04 | X | SPA | @0 Reconocimiento voz @5 04 |
---|
C03 | 05 | X | FRE | @0 Reconnaissance automatique @5 05 |
---|
C03 | 05 | X | ENG | @0 Automatic recognition @5 05 |
---|
C03 | 05 | X | SPA | @0 Reconocimiento automático @5 05 |
---|
C03 | 06 | X | FRE | @0 Filtrage fréquence @5 11 |
---|
C03 | 06 | X | ENG | @0 Frequency filtering @5 11 |
---|
C03 | 06 | X | SPA | @0 Filtrado frecuencia @5 11 |
---|
C03 | 07 | X | FRE | @0 Réponse temporelle @5 13 |
---|
C03 | 07 | X | ENG | @0 Time response @5 13 |
---|
C03 | 07 | X | SPA | @0 Respuesta temporal @5 13 |
---|
C03 | 08 | X | FRE | @0 Cepstre @5 21 |
---|
C03 | 08 | X | ENG | @0 Cepstrum @5 21 |
---|
C03 | 09 | X | FRE | @0 Transformation ondelette @5 24 |
---|
C03 | 09 | X | ENG | @0 Wavelet transformation @5 24 |
---|
C03 | 09 | X | SPA | @0 Transformación ondita @5 24 |
---|
C03 | 10 | X | FRE | @0 4372 @2 PAC @4 INC @5 57 |
---|
N21 | | | | @1 082 |
---|
N82 | | | | @1 PSI |
---|
|
pR |
A30 | 01 | 1 | ENG | @1 ICANN / ICONIP . Joint international conference @3 Istanbul TUR @4 2003-06-26 |
---|
|
Format Inist (serveur)
NO : | PASCAL 04-0130632 INIST |
ET : | Frequency and wavelet filtering for robust speech recognition |
AU : | DEVIREN (Murat); DAOUDI (Khalid); KAYNAK (Okyay); ALPAYDIN (Ethem); OJA (Erkki); LEI XU |
AF : | INRIA-LORIA (Speech Group) B.P. 101/54602 Villers les Nancy/France (1 aut., 2 aut.) |
DT : | Publication en série; Congrès; Niveau analytique |
SO : | Lecture notes in computer science; ISSN 0302-9743; Allemagne; Da. 2003; Vol. 2714; Pp. 452-460; Bibl. 11 ref. |
LA : | Anglais |
EA : | Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case. |
CC : | 001D02C04; 001D04A05B |
FD : | Intelligence artificielle; Interface utilisateur; Base donnée; Reconnaissance parole; Reconnaissance automatique; Filtrage fréquence; Réponse temporelle; Cepstre; Transformation ondelette; 4372 |
ED : | Artificial intelligence; User interface; Database; Speech recognition; Automatic recognition; Frequency filtering; Time response; Cepstrum; Wavelet transformation |
SD : | Inteligencia artificial; Interfase usuario; Base dato; Reconocimiento voz; Reconocimiento automático; Filtrado frecuencia; Respuesta temporal; Transformación ondita |
LO : | INIST-16343.354000117807160540 |
ID : | 04-0130632 |
Links to Exploration step
Pascal:04-0130632
Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" level="a">Frequency and wavelet filtering for robust speech recognition</title>
<author><name sortKey="Deviren, Murat" sort="Deviren, Murat" uniqKey="Deviren M" first="Murat" last="Deviren">Murat Deviren</name>
<affiliation><inist:fA14 i1="01"><s1>INRIA-LORIA (Speech Group) B.P. 101</s1>
<s2>54602 Villers les Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Daoudi, Khalid" sort="Daoudi, Khalid" uniqKey="Daoudi K" first="Khalid" last="Daoudi">Khalid Daoudi</name>
<affiliation><inist:fA14 i1="01"><s1>INRIA-LORIA (Speech Group) B.P. 101</s1>
<s2>54602 Villers les Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">INIST</idno>
<idno type="inist">04-0130632</idno>
<date when="2003">2003</date>
<idno type="stanalyst">PASCAL 04-0130632 INIST</idno>
<idno type="RBID">Pascal:04-0130632</idno>
<idno type="wicri:Area/PascalFrancis/Corpus">000727</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en" level="a">Frequency and wavelet filtering for robust speech recognition</title>
<author><name sortKey="Deviren, Murat" sort="Deviren, Murat" uniqKey="Deviren M" first="Murat" last="Deviren">Murat Deviren</name>
<affiliation><inist:fA14 i1="01"><s1>INRIA-LORIA (Speech Group) B.P. 101</s1>
<s2>54602 Villers les Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
<author><name sortKey="Daoudi, Khalid" sort="Daoudi, Khalid" uniqKey="Daoudi K" first="Khalid" last="Daoudi">Khalid Daoudi</name>
<affiliation><inist:fA14 i1="01"><s1>INRIA-LORIA (Speech Group) B.P. 101</s1>
<s2>54602 Villers les Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</inist:fA14>
</affiliation>
</author>
</analytic>
<series><title level="j" type="main">Lecture notes in computer science</title>
<idno type="ISSN">0302-9743</idno>
<imprint><date when="2003">2003</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
<seriesStmt><title level="j" type="main">Lecture notes in computer science</title>
<idno type="ISSN">0302-9743</idno>
</seriesStmt>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Artificial intelligence</term>
<term>Automatic recognition</term>
<term>Cepstrum</term>
<term>Database</term>
<term>Frequency filtering</term>
<term>Speech recognition</term>
<term>Time response</term>
<term>User interface</term>
<term>Wavelet transformation</term>
</keywords>
<keywords scheme="Pascal" xml:lang="fr"><term>Intelligence artificielle</term>
<term>Interface utilisateur</term>
<term>Base donnée</term>
<term>Reconnaissance parole</term>
<term>Reconnaissance automatique</term>
<term>Filtrage fréquence</term>
<term>Réponse temporelle</term>
<term>Cepstre</term>
<term>Transformation ondelette</term>
<term>4372</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case.</div>
</front>
</TEI>
<inist><standard h6="B"><pA><fA01 i1="01" i2="1"><s0>0302-9743</s0>
</fA01>
<fA05><s2>2714</s2>
</fA05>
<fA08 i1="01" i2="1" l="ENG"><s1>Frequency and wavelet filtering for robust speech recognition</s1>
</fA08>
<fA09 i1="01" i2="1" l="ENG"><s1>Artificial neural networks and neural information processing - ICANN / ICONIP 2003 : Istanbul, 26-29 June 2003</s1>
</fA09>
<fA11 i1="01" i2="1"><s1>DEVIREN (Murat)</s1>
</fA11>
<fA11 i1="02" i2="1"><s1>DAOUDI (Khalid)</s1>
</fA11>
<fA12 i1="01" i2="1"><s1>KAYNAK (Okyay)</s1>
<s9>ed.</s9>
</fA12>
<fA12 i1="02" i2="1"><s1>ALPAYDIN (Ethem)</s1>
<s9>ed.</s9>
</fA12>
<fA12 i1="03" i2="1"><s1>OJA (Erkki)</s1>
<s9>ed.</s9>
</fA12>
<fA12 i1="04" i2="1"><s1>LEI XU</s1>
<s9>ed.</s9>
</fA12>
<fA14 i1="01"><s1>INRIA-LORIA (Speech Group) B.P. 101</s1>
<s2>54602 Villers les Nancy</s2>
<s3>FRA</s3>
<sZ>1 aut.</sZ>
<sZ>2 aut.</sZ>
</fA14>
<fA20><s1>452-460</s1>
</fA20>
<fA21><s1>2003</s1>
</fA21>
<fA23 i1="01"><s0>ENG</s0>
</fA23>
<fA26 i1="01"><s0>3-540-40408-2</s0>
</fA26>
<fA43 i1="01"><s1>INIST</s1>
<s2>16343</s2>
<s5>354000117807160540</s5>
</fA43>
<fA44><s0>0000</s0>
<s1>© 2004 INIST-CNRS. All rights reserved.</s1>
</fA44>
<fA45><s0>11 ref.</s0>
</fA45>
<fA47 i1="01" i2="1"><s0>04-0130632</s0>
</fA47>
<fA60><s1>P</s1>
<s2>C</s2>
</fA60>
<fA64 i1="01" i2="1"><s0>Lecture notes in computer science</s0>
</fA64>
<fA66 i1="01"><s0>DEU</s0>
</fA66>
<fC01 i1="01" l="ENG"><s0>Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case.</s0>
</fC01>
<fC02 i1="01" i2="X"><s0>001D02C04</s0>
</fC02>
<fC02 i1="02" i2="X"><s0>001D04A05B</s0>
</fC02>
<fC03 i1="01" i2="X" l="FRE"><s0>Intelligence artificielle</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="ENG"><s0>Artificial intelligence</s0>
<s5>01</s5>
</fC03>
<fC03 i1="01" i2="X" l="SPA"><s0>Inteligencia artificial</s0>
<s5>01</s5>
</fC03>
<fC03 i1="02" i2="X" l="FRE"><s0>Interface utilisateur</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="ENG"><s0>User interface</s0>
<s5>02</s5>
</fC03>
<fC03 i1="02" i2="X" l="SPA"><s0>Interfase usuario</s0>
<s5>02</s5>
</fC03>
<fC03 i1="03" i2="X" l="FRE"><s0>Base donnée</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="ENG"><s0>Database</s0>
<s5>03</s5>
</fC03>
<fC03 i1="03" i2="X" l="SPA"><s0>Base dato</s0>
<s5>03</s5>
</fC03>
<fC03 i1="04" i2="X" l="FRE"><s0>Reconnaissance parole</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="ENG"><s0>Speech recognition</s0>
<s5>04</s5>
</fC03>
<fC03 i1="04" i2="X" l="SPA"><s0>Reconocimiento voz</s0>
<s5>04</s5>
</fC03>
<fC03 i1="05" i2="X" l="FRE"><s0>Reconnaissance automatique</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="ENG"><s0>Automatic recognition</s0>
<s5>05</s5>
</fC03>
<fC03 i1="05" i2="X" l="SPA"><s0>Reconocimiento automático</s0>
<s5>05</s5>
</fC03>
<fC03 i1="06" i2="X" l="FRE"><s0>Filtrage fréquence</s0>
<s5>11</s5>
</fC03>
<fC03 i1="06" i2="X" l="ENG"><s0>Frequency filtering</s0>
<s5>11</s5>
</fC03>
<fC03 i1="06" i2="X" l="SPA"><s0>Filtrado frecuencia</s0>
<s5>11</s5>
</fC03>
<fC03 i1="07" i2="X" l="FRE"><s0>Réponse temporelle</s0>
<s5>13</s5>
</fC03>
<fC03 i1="07" i2="X" l="ENG"><s0>Time response</s0>
<s5>13</s5>
</fC03>
<fC03 i1="07" i2="X" l="SPA"><s0>Respuesta temporal</s0>
<s5>13</s5>
</fC03>
<fC03 i1="08" i2="X" l="FRE"><s0>Cepstre</s0>
<s5>21</s5>
</fC03>
<fC03 i1="08" i2="X" l="ENG"><s0>Cepstrum</s0>
<s5>21</s5>
</fC03>
<fC03 i1="09" i2="X" l="FRE"><s0>Transformation ondelette</s0>
<s5>24</s5>
</fC03>
<fC03 i1="09" i2="X" l="ENG"><s0>Wavelet transformation</s0>
<s5>24</s5>
</fC03>
<fC03 i1="09" i2="X" l="SPA"><s0>Transformación ondita</s0>
<s5>24</s5>
</fC03>
<fC03 i1="10" i2="X" l="FRE"><s0>4372</s0>
<s2>PAC</s2>
<s4>INC</s4>
<s5>57</s5>
</fC03>
<fN21><s1>082</s1>
</fN21>
<fN82><s1>PSI</s1>
</fN82>
</pA>
<pR><fA30 i1="01" i2="1" l="ENG"><s1>ICANN / ICONIP . Joint international conference</s1>
<s3>Istanbul TUR</s3>
<s4>2003-06-26</s4>
</fA30>
</pR>
</standard>
<server><NO>PASCAL 04-0130632 INIST</NO>
<ET>Frequency and wavelet filtering for robust speech recognition</ET>
<AU>DEVIREN (Murat); DAOUDI (Khalid); KAYNAK (Okyay); ALPAYDIN (Ethem); OJA (Erkki); LEI XU</AU>
<AF>INRIA-LORIA (Speech Group) B.P. 101/54602 Villers les Nancy/France (1 aut., 2 aut.)</AF>
<DT>Publication en série; Congrès; Niveau analytique</DT>
<SO>Lecture notes in computer science; ISSN 0302-9743; Allemagne; Da. 2003; Vol. 2714; Pp. 452-460; Bibl. 11 ref.</SO>
<LA>Anglais</LA>
<EA>Mel-frequency cepstral coefficients (MFCC) are the most widely used features in current speech recognition systems. However, they have a poor physical interpretation and they do not lie in the frequency domain. Frequency filtering (FF) is a technique that has been recently developed to design frequency-localized speech features that perform similar to MFCC in terms of recognition performances. Motivated by our desire to build time-frequency speech models, we wanted to use the FF technique as front-end. However, when evaluating FF on the Aurora-3 database we found some discrepancies in the highly mismatch case. This led us to put FF in another perspective: the wavelet transform. By doing so, we were able to explain the discrepancies and to achieve significant improvements in recognition in the highly mismatch case.</EA>
<CC>001D02C04; 001D04A05B</CC>
<FD>Intelligence artificielle; Interface utilisateur; Base donnée; Reconnaissance parole; Reconnaissance automatique; Filtrage fréquence; Réponse temporelle; Cepstre; Transformation ondelette; 4372</FD>
<ED>Artificial intelligence; User interface; Database; Speech recognition; Automatic recognition; Frequency filtering; Time response; Cepstrum; Wavelet transformation</ED>
<SD>Inteligencia artificial; Interfase usuario; Base dato; Reconocimiento voz; Reconocimiento automático; Filtrado frecuencia; Respuesta temporal; Transformación ondita</SD>
<LO>INIST-16343.354000117807160540</LO>
<ID>04-0130632</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 000727 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PascalFrancis/Corpus/biblio.hfd -nk 000727 | 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:04-0130632
|texte= Frequency and wavelet filtering for robust 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 | ![](Common/icons/LogoDilib.gif) |