Speech Recognition based on Second Order HMM
Identifieur interne : 00E214 ( Main/Merge ); précédent : 00E213; suivant : 00E215Speech Recognition based on Second Order HMM
Auteurs : A. Kriouile ; J.-F. Mari ; Jean-Paul Haton [France]Source :
English descriptors
Abstract
Automatic speech recognition plays an important role in the framework of man-machine communication. Substantial industrial developments are at present in progress in this area. However, several fundamental questions remain open. This paper presents a new recognizer for isolated words, based on second order Hidden Markov models (HMM). We propose a formulation of the Baum-Welch algorithm and an extension of the Viterbi algorithm that make such second order models computationally efficient for real-time applications. A comparative study between first order and second order systems is carried out. To evaluate the performances of both systems, we have used a database of french digits. It can be seen that the results are more accurate with recognizer using second order HMM. In a speaker-independent mode, the performance increase from 91========percnt; (first order) to 93========percnt; (second order) in the same experimental conditions. In a multi-speaker mode, recognition accuracy increased from 97========percnt; to 99========percnt;.
Links toward previous steps (curation, corpus...)
- to stream Crin, to step Corpus: 000C51
- to stream Crin, to step Curation: 000C51
- to stream Crin, to step Checkpoint: 003A13
Links to Exploration step
CRIN:kriouile91aLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="215">Speech Recognition based on Second Order HMM</title></titleStmt><publicationStmt><idno type="RBID">CRIN:kriouile91a</idno><date when="1991" year="1991">1991</date><idno type="wicri:Area/Crin/Corpus">000C51</idno><idno type="wicri:Area/Crin/Curation">000C51</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Curation">000C51</idno><idno type="wicri:Area/Crin/Checkpoint">003A13</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Checkpoint">003A13</idno><idno type="wicri:Area/Main/Merge">00E214</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">Speech Recognition based on Second Order HMM</title><author><name sortKey="Kriouile, A" sort="Kriouile, A" uniqKey="Kriouile A" first="A." last="Kriouile">A. Kriouile</name></author><author><name sortKey="Mari, J F" sort="Mari, J F" uniqKey="Mari J" first="J.-F." last="Mari">J.-F. Mari</name></author><author><name sortKey="Haton, J P" sort="Haton, J P" uniqKey="Haton J" first="J.-P." last="Haton">Jean-Paul Haton</name><affiliation><country>France</country><placeName><settlement type="city">Nancy</settlement><region type="region" nuts="2">Grand Est</region><region type="region" nuts="2">Lorraine (région)</region></placeName><orgName type="laboratoire" n="5">Laboratoire lorrain de recherche en informatique et ses applications</orgName><orgName type="university">Université de Lorraine</orgName><orgName type="institution">Centre national de la recherche scientifique</orgName><orgName type="institution">Institut national de recherche en informatique et en automatique</orgName></affiliation></author></analytic></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>man-machine communication</term><term>speech recognition</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en" wicri:score="1878">Automatic speech recognition plays an important role in the framework of man-machine communication. Substantial industrial developments are at present in progress in this area. However, several fundamental questions remain open. This paper presents a new recognizer for isolated words, based on second order Hidden Markov models (HMM). We propose a formulation of the Baum-Welch algorithm and an extension of the Viterbi algorithm that make such second order models computationally efficient for real-time applications. A comparative study between first order and second order systems is carried out. To evaluate the performances of both systems, we have used a database of french digits. It can be seen that the results are more accurate with recognizer using second order HMM. In a speaker-independent mode, the performance increase from 91========percnt; (first order) to 93========percnt; (second order) in the same experimental conditions. In a multi-speaker mode, recognition accuracy increased from 97========percnt; to 99========percnt;.</div></front></TEI></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Main/Merge
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 00E214 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Merge/biblio.hfd -nk 00E214 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Lorraine
|area= InforLorV4
|flux= Main
|étape= Merge
|type= RBID
|clé= CRIN:kriouile91a
|texte= Speech Recognition based on Second Order HMM
}}
| This area was generated with Dilib version V0.6.33. |  |