Classifier Acceleration by Imitation
Identifieur interne : 000497 ( Main/Exploration ); précédent : 000496; suivant : 000498Classifier Acceleration by Imitation
Auteurs : Takahiro Ota [Japon] ; Toshikazu Wada [Japon] ; Takayuki Nakamura [Japon]Source :
- Lecture Notes in Computer Science [ 0302-9743 ] ; 2011.
Abstract
Abstract: This paper presents a framework named “Classifier Molding” that imitates arbitrary classifiers by linear regression trees so as to accelerate classification speed. This framework requires an accurate (but slow) classifier and large amount of training data. As an example of accurate classifier, we used the Compound Similarity Method (CSM) for Industrial Ink Jet Printer (IIJP) character recognition problem. The input-output relationship of trained CSM is imitated by a linear regression tree by providing a large amount of training data. For generating the training data, we developed a character pattern fluctuation method simulating the IIJP printing process. The learnt linear regression tree can be used as an accelerated classifier. Based on this classifier, we also developed Classification based Character Segmentation (CCS) method, which extracts character patterns from an image so as to maximize the total classification scores. Through extensive experiments, we confirmed that imitated classifiers are 1500 times faster than the original classifier without dropping the recognition rate and CCS method greatly corrects the segmentation errors of bottom-up segmentation method.
Url:
DOI: 10.1007/978-3-642-19282-1_52
Affiliations:
Links toward previous steps (curation, corpus...)
- to stream Istex, to step Corpus: 000208
- to stream Istex, to step Curation: 000205
- to stream Istex, to step Checkpoint: 000154
- to stream Main, to step Merge: 000503
- to stream Main, to step Curation: 000497
Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">Classifier Acceleration by Imitation</title><author><name sortKey="Ota, Takahiro" sort="Ota, Takahiro" uniqKey="Ota T" first="Takahiro" last="Ota">Takahiro Ota</name></author><author><name sortKey="Wada, Toshikazu" sort="Wada, Toshikazu" uniqKey="Wada T" first="Toshikazu" last="Wada">Toshikazu Wada</name></author><author><name sortKey="Nakamura, Takayuki" sort="Nakamura, Takayuki" uniqKey="Nakamura T" first="Takayuki" last="Nakamura">Takayuki Nakamura</name></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:1A45CF46D682EE73B5F57C307B8DDBBCEEB770CC</idno><date when="2011" year="2011">2011</date><idno type="doi">10.1007/978-3-642-19282-1_52</idno><idno type="url">https://api.istex.fr/document/1A45CF46D682EE73B5F57C307B8DDBBCEEB770CC/fulltext/pdf</idno><idno type="wicri:Area/Istex/Corpus">000208</idno><idno type="wicri:Area/Istex/Curation">000205</idno><idno type="wicri:Area/Istex/Checkpoint">000154</idno><idno type="wicri:doubleKey">0302-9743:2011:Ota T:classifier:acceleration:by</idno><idno type="wicri:Area/Main/Merge">000503</idno><idno type="wicri:Area/Main/Curation">000497</idno><idno type="wicri:Area/Main/Exploration">000497</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">Classifier Acceleration by Imitation</title><author><name sortKey="Ota, Takahiro" sort="Ota, Takahiro" uniqKey="Ota T" first="Takahiro" last="Ota">Takahiro Ota</name><affiliation wicri:level="1"><country xml:lang="fr">Japon</country><wicri:regionArea>Wakayama University, No. 930 Sakaedani, Wakayama-shi, 640-8510, Wakayama</wicri:regionArea><wicri:noRegion>Wakayama</wicri:noRegion></affiliation><affiliation wicri:level="1"><country xml:lang="fr">Japon</country><wicri:regionArea>Kishugikenkogyo Co., Ltd, No. 466 Nunobiki, Wakayama-shi, 641-0015, Wakayama</wicri:regionArea><wicri:noRegion>Wakayama</wicri:noRegion></affiliation><affiliation wicri:level="1"><country wicri:rule="url">Japon</country></affiliation></author><author><name sortKey="Wada, Toshikazu" sort="Wada, Toshikazu" uniqKey="Wada T" first="Toshikazu" last="Wada">Toshikazu Wada</name><affiliation wicri:level="1"><country xml:lang="fr">Japon</country><wicri:regionArea>Wakayama University, No. 930 Sakaedani, Wakayama-shi, 640-8510, Wakayama</wicri:regionArea><wicri:noRegion>Wakayama</wicri:noRegion></affiliation><affiliation><wicri:noCountry code="no comma">E-mail: twada@ieee.org</wicri:noCountry></affiliation></author><author><name sortKey="Nakamura, Takayuki" sort="Nakamura, Takayuki" uniqKey="Nakamura T" first="Takayuki" last="Nakamura">Takayuki Nakamura</name><affiliation wicri:level="1"><country xml:lang="fr">Japon</country><wicri:regionArea>Wakayama University, No. 930 Sakaedani, Wakayama-shi, 640-8510, Wakayama</wicri:regionArea><wicri:noRegion>Wakayama</wicri:noRegion></affiliation><affiliation wicri:level="1"><country wicri:rule="url">Japon</country></affiliation></author></analytic><monogr></monogr><series><title level="s">Lecture Notes in Computer Science</title><imprint><date>2011</date></imprint><idno type="ISSN">0302-9743</idno><idno type="eISSN">1611-3349</idno><idno type="ISSN">0302-9743</idno></series><idno type="istex">1A45CF46D682EE73B5F57C307B8DDBBCEEB770CC</idno><idno type="DOI">10.1007/978-3-642-19282-1_52</idno><idno type="ChapterID">52</idno><idno type="ChapterID">Chap52</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: This paper presents a framework named “Classifier Molding” that imitates arbitrary classifiers by linear regression trees so as to accelerate classification speed. This framework requires an accurate (but slow) classifier and large amount of training data. As an example of accurate classifier, we used the Compound Similarity Method (CSM) for Industrial Ink Jet Printer (IIJP) character recognition problem. The input-output relationship of trained CSM is imitated by a linear regression tree by providing a large amount of training data. For generating the training data, we developed a character pattern fluctuation method simulating the IIJP printing process. The learnt linear regression tree can be used as an accelerated classifier. Based on this classifier, we also developed Classification based Character Segmentation (CCS) method, which extracts character patterns from an image so as to maximize the total classification scores. Through extensive experiments, we confirmed that imitated classifiers are 1500 times faster than the original classifier without dropping the recognition rate and CCS method greatly corrects the segmentation errors of bottom-up segmentation method.</div></front></TEI><affiliations><list><country><li>Japon</li></country></list><tree><country name="Japon"><noRegion><name sortKey="Ota, Takahiro" sort="Ota, Takahiro" uniqKey="Ota T" first="Takahiro" last="Ota">Takahiro Ota</name></noRegion><name sortKey="Nakamura, Takayuki" sort="Nakamura, Takayuki" uniqKey="Nakamura T" first="Takayuki" last="Nakamura">Takayuki Nakamura</name><name sortKey="Nakamura, Takayuki" sort="Nakamura, Takayuki" uniqKey="Nakamura T" first="Takayuki" last="Nakamura">Takayuki Nakamura</name><name sortKey="Ota, Takahiro" sort="Ota, Takahiro" uniqKey="Ota T" first="Takahiro" last="Ota">Takahiro Ota</name><name sortKey="Ota, Takahiro" sort="Ota, Takahiro" uniqKey="Ota T" first="Takahiro" last="Ota">Takahiro Ota</name><name sortKey="Wada, Toshikazu" sort="Wada, Toshikazu" uniqKey="Wada T" first="Toshikazu" last="Wada">Toshikazu Wada</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 000497 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Main/Exploration/biblio.hfd -nk 000497 | 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:1A45CF46D682EE73B5F57C307B8DDBBCEEB770CC
|texte= Classifier Acceleration by Imitation
}}
| This area was generated with Dilib version V0.6.32. |  |