Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming
Identifieur interne : 000C91 ( Crin/Corpus ); précédent : 000C90; suivant : 000C92Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming
Auteurs : Y. Gong ; J.-P. HatonSource :
- IEEE Transactions on Pattern Analysis and Machine Intelligence ; 1991.
English descriptors
Abstract
Many pattern recognition problems involve signal-to-string conversion. The string recognition problem can be formulated as the maximization of a constrained time integral of a sequence of likelihood functions. Such likelihood functions are time series of likelihood ratios between image of component symbols and input data. We propose in this paper a new method of conversion based on embedded dynamic programming which can adapt its search to the variation of the input signal. The optimizing process is guided by high-valued portions of likelihood function of symbols composing the string and is solved by two embedded dynamic programming process. Applied to continuous speech recognition using phoneme as basic recognition unit on a 100-word vocabulary, the method achieved 4========percnt; improvement of recognition rate in a 1/20 time compared to a classical DP-based method.
Links to Exploration step
CRIN:gong91cLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="870">Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming</title>
</titleStmt>
<publicationStmt><idno type="RBID">CRIN:gong91c</idno>
<date when="1991" year="1991">1991</date>
<idno type="wicri:Area/Crin/Corpus">000C91</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en">Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming</title>
<author><name sortKey="Gong, Y" sort="Gong, Y" uniqKey="Gong Y" first="Y." last="Gong">Y. Gong</name>
</author>
<author><name sortKey="Haton, J P" sort="Haton, J P" uniqKey="Haton J" first="J.-P." last="Haton">J.-P. Haton</name>
</author>
</analytic>
<series><title level="j">IEEE Transactions on Pattern Analysis and Machine Intelligence</title>
<imprint><date when="1991" type="published">1991</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>dynamic programming</term>
<term>likelihood function</term>
<term>speech recognition</term>
<term>string recognition</term>
</keywords>
</textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en" wicri:score="4933">Many pattern recognition problems involve signal-to-string conversion. The string recognition problem can be formulated as the maximization of a constrained time integral of a sequence of likelihood functions. Such likelihood functions are time series of likelihood ratios between image of component symbols and input data. We propose in this paper a new method of conversion based on embedded dynamic programming which can adapt its search to the variation of the input signal. The optimizing process is guided by high-valued portions of likelihood function of symbols composing the string and is solved by two embedded dynamic programming process. Applied to continuous speech recognition using phoneme as basic recognition unit on a 100-word vocabulary, the method achieved 4========percnt; improvement of recognition rate in a 1/20 time compared to a classical DP-based method.</div>
</front>
</TEI>
<BibTex type="article"><ref>gong91c</ref>
<crinnumber>91-R-243</crinnumber>
<category>1</category>
<equipe>RFIA</equipe>
<author><e>Gong, Y.</e>
<e>Haton, J.-P.</e>
</author>
<title>Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming</title>
<journal>IEEE Transactions on Pattern Analysis and Machine Intelligence</journal>
<year>1991</year>
<volume>13</volume>
<number>3</number>
<pages>297-302</pages>
<month>Mar</month>
<keywords><e>dynamic programming</e>
<e>likelihood function</e>
<e>speech recognition</e>
<e>string recognition</e>
</keywords>
<abstract>Many pattern recognition problems involve signal-to-string conversion. The string recognition problem can be formulated as the maximization of a constrained time integral of a sequence of likelihood functions. Such likelihood functions are time series of likelihood ratios between image of component symbols and input data. We propose in this paper a new method of conversion based on embedded dynamic programming which can adapt its search to the variation of the input signal. The optimizing process is guided by high-valued portions of likelihood function of symbols composing the string and is solved by two embedded dynamic programming process. Applied to continuous speech recognition using phoneme as basic recognition unit on a 100-word vocabulary, the method achieved 4========percnt; improvement of recognition rate in a 1/20 time compared to a classical DP-based method.</abstract>
</BibTex>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Crin/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000C91 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Crin/Corpus/biblio.hfd -nk 000C91 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Wicri/Lorraine |area= InforLorV4 |flux= Crin |étape= Corpus |type= RBID |clé= CRIN:gong91c |texte= Signal-to-String Conversion Based on High Likelihood Regions Using Embedded Dynamic Programming }}
![]() | This area was generated with Dilib version V0.6.33. | ![]() |