Combining Multiple Classifiers for Faster Optical Character Recognition
Identifieur interne : 001099 ( Main/Exploration ); précédent : 001098; suivant : 001100Combining Multiple Classifiers for Faster Optical Character Recognition
Auteurs : Kumar Chellapilla [États-Unis] ; Michael Shilman [États-Unis] ; Patrice Simard [États-Unis]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2006.
Abstract
Abstract: Traditional approaches to combining classifiers attempt to improve classification accuracy at the cost of increased processing. They may be viewed as providing an accuracy-speed trade-off: higher accuracy for lower speed. In this paper we present a novel approach to combining multiple classifiers to solve the inverse problem of significantly improving classification speeds at the cost of slightly reduced classification accuracy. We propose a cascade architecture for combining classifiers and cast the process of building such a cascade as a search and optimization problem. We present two algorithms based on steepest-descent and dynamic programming for producing approximate solutions fast. We also present a simulated annealing algorithm and a depth-first-search algorithm for finding optimal solutions. Results on handwritten optical character recognition indicate that a) a speedup of 4-9 times is possible with no increase in error and b) speedups of up to 15 times are possible when twice as many errors can be tolerated.
Url:
DOI: 10.1007/11669487_32
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 000730
- to stream Istex, to step Curation: 000722
- to stream Istex, to step Checkpoint: 000A72
- to stream Main, to step Merge: 001116
- to stream Main, to step Curation: 001099
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Combining Multiple Classifiers for Faster Optical Character Recognition</title><author><name sortKey="Chellapilla, Kumar" sort="Chellapilla, Kumar" uniqKey="Chellapilla K" first="Kumar" last="Chellapilla">Kumar Chellapilla</name></author><author><name sortKey="Shilman, Michael" sort="Shilman, Michael" uniqKey="Shilman M" first="Michael" last="Shilman">Michael Shilman</name></author><author><name sortKey="Simard, Patrice" sort="Simard, Patrice" uniqKey="Simard P" first="Patrice" last="Simard">Patrice Simard</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:7545EC2BA473A52004C6FA4B7645B3D4398A31DD</idno><date when="2006" year="2006">2006</date><idno type="doi">10.1007/11669487_32</idno><idno type="url">https://api.istex.fr/document/7545EC2BA473A52004C6FA4B7645B3D4398A31DD/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000730</idno><idno type="wicri:Area/Istex/Curation">000722</idno><idno type="wicri:Area/Istex/Checkpoint">000A72</idno><idno type="wicri:doubleKey">0302-9743:2006:Chellapilla K:combining:multiple:classifiers</idno><idno type="wicri:Area/Main/Merge">001116</idno><idno type="wicri:Area/Main/Curation">001099</idno><idno type="wicri:Area/Main/Exploration">001099</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Combining Multiple Classifiers for Faster Optical Character Recognition</title><author><name sortKey="Chellapilla, Kumar" sort="Chellapilla, Kumar" uniqKey="Chellapilla K" first="Kumar" last="Chellapilla">Kumar Chellapilla</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Microsoft Research, One Microsoft Way, 98052, Redmond, WA</wicri:regionArea><placeName><region type="state">Washington (État)</region></placeName></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: kumarc@microsoft.com</wicri:noCountry></affiliation></author><author><name sortKey="Shilman, Michael" sort="Shilman, Michael" uniqKey="Shilman M" first="Michael" last="Shilman">Michael Shilman</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Microsoft Research, One Microsoft Way, 98052, Redmond, WA</wicri:regionArea><placeName><region type="state">Washington (État)</region></placeName></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: shilman@microsoft.com</wicri:noCountry></affiliation></author><author><name sortKey="Simard, Patrice" sort="Simard, Patrice" uniqKey="Simard P" first="Patrice" last="Simard">Patrice Simard</name><affiliation wicri:level="2"><country xml:lang="fr">États-Unis</country><wicri:regionArea>Microsoft Research, One Microsoft Way, 98052, Redmond, WA</wicri:regionArea><placeName><region type="state">Washington (État)</region></placeName></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: patrice@microsoft.com</wicri:noCountry></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2006</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">7545EC2BA473A52004C6FA4B7645B3D4398A31DD</idno><idno type="DOI">10.1007/11669487_32</idno><idno type="ChapterID">32</idno><idno type="ChapterID">Chap32</idno></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">0302-9743</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass><langUsage><language ident="en">en</language></langUsage></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Abstract: Traditional approaches to combining classifiers attempt to improve classification accuracy at the cost of increased processing. They may be viewed as providing an accuracy-speed trade-off: higher accuracy for lower speed. In this paper we present a novel approach to combining multiple classifiers to solve the inverse problem of significantly improving classification speeds at the cost of slightly reduced classification accuracy. We propose a cascade architecture for combining classifiers and cast the process of building such a cascade as a search and optimization problem. We present two algorithms based on steepest-descent and dynamic programming for producing approximate solutions fast. We also present a simulated annealing algorithm and a depth-first-search algorithm for finding optimal solutions. Results on handwritten optical character recognition indicate that a) a speedup of 4-9 times is possible with no increase in error and b) speedups of up to 15 times are possible when twice as many errors can be tolerated.</div></front></TEI><affiliations><list><country><li>États-Unis</li></country><region><li>Washington (État)</li></region></list><tree><country name="États-Unis"><region name="Washington (État)"><name sortKey="Chellapilla, Kumar" sort="Chellapilla, Kumar" uniqKey="Chellapilla K" first="Kumar" last="Chellapilla">Kumar Chellapilla</name></region><name sortKey="Shilman, Michael" sort="Shilman, Michael" uniqKey="Shilman M" first="Michael" last="Shilman">Michael Shilman</name><name sortKey="Simard, Patrice" sort="Simard, Patrice" uniqKey="Simard P" first="Patrice" last="Simard">Patrice Simard</name></country></tree></affiliations></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/Main/Exploration
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 001099 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 001099 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Ticri/CIDE
|area= OcrV1
|flux= Main
|étape= Exploration
|type= RBID
|clé= ISTEX:7545EC2BA473A52004C6FA4B7645B3D4398A31DD
|texte= Combining Multiple Classifiers for Faster Optical Character Recognition
}}
| This area was generated with Dilib version V0.6.32. |  |