Use of Stochastic Models in Text Recognition
Identifieur interne : 00D380 ( Main/Merge ); précédent : 00D379; suivant : 00D381Use of Stochastic Models in Text Recognition
Auteurs : Abdel Belaïd [France] ; G. SaonSource :
English descriptors
Abstract
We present in this paper two applications of stochastic models to text recognition. The first application concerns multi-font printed text recognition (PTR) while the second deals with handwritten word recognition (HWR). The former is built around first and second order hidden Markov models (HMM) and uses an extended Viterbi algorithm for recognition. The method operates in a bottom-up manner by proposing a list of candidates for each character and then the system uses combinations of stochastic and dictionary verification methods for word recognition and error-correction. The later deals with unconstrained off-line HWR with a limited vocabulary. It is based on the maximization of a given word in terms of probabilities of its component characters. This approach operates in a top-down manner by giving for each word of the lexicon against which the pattern is matched the character plausibility. The word having the highest average plausibility per character is selected. This approach tends to bypass the difficult problem of segmentation.
Links toward previous steps (curation, corpus...)
- to stream Crin, to step Corpus: 001498
- to stream Crin, to step Curation: 001498
- to stream Crin, to step Checkpoint: 003082
Links to Exploration step
CRIN:belaid94fLe document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en" wicri:score="271">Use of Stochastic Models in Text Recognition</title></titleStmt><publicationStmt><idno type="RBID">CRIN:belaid94f</idno><date when="1994" year="1994">1994</date><idno type="wicri:Area/Crin/Corpus">001498</idno><idno type="wicri:Area/Crin/Curation">001498</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Curation">001498</idno><idno type="wicri:Area/Crin/Checkpoint">003082</idno><idno type="wicri:explorRef" wicri:stream="Crin" wicri:step="Checkpoint">003082</idno><idno type="wicri:Area/Main/Merge">00D380</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">Use of Stochastic Models in Text Recognition</title><author><name sortKey="Belaid, A" sort="Belaid, A" uniqKey="Belaid A" first="A." last="Belaïd">Abdel Belaïd</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><author><name sortKey="Saon, G" sort="Saon, G" uniqKey="Saon G" first="G." last="Saon">G. Saon</name></author></analytic></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Hidden Markov models</term><term>statistic correlation</term><term>text recognition</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en" wicri:score="4156">We present in this paper two applications of stochastic models to text recognition. The first application concerns multi-font printed text recognition (PTR) while the second deals with handwritten word recognition (HWR). The former is built around first and second order hidden Markov models (HMM) and uses an extended Viterbi algorithm for recognition. The method operates in a bottom-up manner by proposing a list of candidates for each character and then the system uses combinations of stochastic and dictionary verification methods for word recognition and error-correction. The later deals with unconstrained off-line HWR with a limited vocabulary. It is based on the maximization of a given word in terms of probabilities of its component characters. This approach operates in a top-down manner by giving for each word of the lexicon against which the pattern is matched the character plausibility. The word having the highest average plausibility per character is selected. This approach tends to bypass the difficult problem of segmentation.</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 00D380 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Merge/biblio.hfd -nk 00D380 | 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:belaid94f
|texte= Use of Stochastic Models in Text Recognition
}}
| This area was generated with Dilib version V0.6.33. |  |