A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology
Identifieur interne : 003023 ( Istex/Corpus ); précédent : 003022; suivant : 003024A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology
Auteurs : Hubert Cecotti ; Abdel BelaïdSource :
- Advances in Soft Computing [ 1615-3871 ]
Abstract
Abstract: We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.
Url:
DOI: 10.1007/3-540-32390-2_13
Links to Exploration step
ISTEX:CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9Le document en format XML
<record><TEI wicri:istexFullTextTei="biblStruct"><teiHeader><fileDesc><titleStmt><title xml:lang="en">A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title><author><name sortKey="Cecotti, Hubert" sort="Cecotti, Hubert" uniqKey="Cecotti H" first="Hubert" last="Cecotti">Hubert Cecotti</name><affiliation><mods:affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: hubert.cecotti@loria.fr</mods:affiliation></affiliation></author><author><name sortKey="Belaid, Abdel" sort="Belaid, Abdel" uniqKey="Belaid A" first="Abdel" last="Belaïd">Abdel Belaïd</name><affiliation><mods:affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: abdel.belaid@loria.fr</mods:affiliation></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">ISTEX</idno><idno type="RBID">ISTEX:CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9</idno><date when="2005" year="2005">2005</date><idno type="doi">10.1007/3-540-32390-2_13</idno><idno type="url">https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/fulltext.pdf</idno><idno type="wicri:Area/Istex/Corpus">003023</idno><idno type="wicri:explorRef" wicri:stream="Istex" wicri:step="Corpus" wicri:corpus="ISTEX">003023</idno></publicationStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title><author><name sortKey="Cecotti, Hubert" sort="Cecotti, Hubert" uniqKey="Cecotti H" first="Hubert" last="Cecotti">Hubert Cecotti</name><affiliation><mods:affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: hubert.cecotti@loria.fr</mods:affiliation></affiliation></author><author><name sortKey="Belaid, Abdel" sort="Belaid, Abdel" uniqKey="Belaid A" first="Abdel" last="Belaïd">Abdel Belaïd</name><affiliation><mods:affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</mods:affiliation></affiliation><affiliation><mods:affiliation>E-mail: abdel.belaid@loria.fr</mods:affiliation></affiliation></author></analytic><monogr></monogr><series><title level="s" type="main" xml:lang="en">Advances in Soft Computing</title><idno type="ISSN">1615-3871</idno><idno type="ISSN">1615-3871</idno></series></biblStruct></sourceDesc><seriesStmt><idno type="ISSN">1615-3871</idno></seriesStmt></fileDesc><profileDesc><textClass></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Abstract: We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.</div></front></TEI><istex><corpusName>springer-ebooks</corpusName><author><json:item><name>Hubert Cecotti</name><affiliations><json:string>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</json:string><json:string>E-mail: hubert.cecotti@loria.fr</json:string></affiliations></json:item><json:item><name>Abdel Belaïd</name><affiliations><json:string>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</json:string><json:string>E-mail: abdel.belaid@loria.fr</json:string></affiliations></json:item></author><arkIstex>ark:/67375/HCB-WMVS30DS-M</arkIstex><language><json:string>eng</json:string></language><originalGenre><json:string>OriginalPaper</json:string></originalGenre><abstract>Abstract: We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.</abstract><qualityIndicators><score>6.869</score><pdfWordCount>2853</pdfWordCount><pdfCharCount>15625</pdfCharCount><pdfVersion>1.3</pdfVersion><pdfPageCount>8</pdfPageCount><pdfPageSize>439.37 x 666.142 pts</pdfPageSize><refBibsNative>false</refBibsNative><abstractWordCount>168</abstractWordCount><abstractCharCount>1094</abstractCharCount><keywordCount>0</keywordCount></qualityIndicators><title>A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title><chapterId><json:string>13</json:string><json:string>Chap13</json:string></chapterId><genre><json:string>conference</json:string></genre><serie><title>Advances in Soft Computing</title><language><json:string>unknown</json:string></language><copyrightDate>2005</copyrightDate><issn><json:string>1615-3871</json:string></issn><volume>Part II</volume><editor><json:item><name>Prof. Janusz Kacprzyk</name><affiliations><json:string>Systems Research Institute, Polish Academy of Sciences, ul. Newelska 6, 01-447, Warsaw, Poland</json:string><json:string>E-mail: kacprzyk@ibspan.waw.pl</json:string></affiliations></json:item></editor></serie><host><title>Computer Recognition Systems</title><language><json:string>unknown</json:string></language><copyrightDate>2005</copyrightDate><doi><json:string>10.1007/3-540-32390-2</json:string></doi><issn><json:string>1615-3871</json:string></issn><eisbn><json:string>978-3-540-32390-7</json:string></eisbn><bookId><json:string>978-3-540-32390-7</json:string></bookId><isbn><json:string>978-3-540-25054-8</json:string></isbn><volume>30</volume><pages><first>129</first><last>136</last></pages><genre><json:string>book-series</json:string></genre><editor><json:item><name>Prof. Marek Kurzyński</name><affiliations><json:string>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</json:string></affiliations></json:item><json:item><name>Dr. Edward Puchała</name><affiliations><json:string>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</json:string></affiliations></json:item><json:item><name>Dr. Michał Woźniak</name><affiliations><json:string>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</json:string></affiliations></json:item><json:item><name>Dr. Andrzej żołnierek</name><affiliations><json:string>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</json:string></affiliations></json:item></editor><subject><json:item><value>Engineering</value></json:item><json:item><value>Computer Science</value></json:item><json:item><value>Pattern Recognition</value></json:item><json:item><value>Artificial Intelligence (incl. Robotics)</value></json:item><json:item><value>Appl.Mathematics/Computational Methods of Engineering</value></json:item><json:item><value>Applications of Mathematics</value></json:item><json:item><value>Information Systems and Communication Service</value></json:item></subject></host><ark><json:string>ark:/67375/HCB-WMVS30DS-M</json:string></ark><publicationDate>2005</publicationDate><copyrightDate>2005</copyrightDate><doi><json:string>10.1007/3-540-32390-2_13</json:string></doi><id>CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9</id><score>1</score><fulltext><json:item><extension>pdf</extension><original>true</original><mimetype>application/pdf</mimetype><uri>https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/fulltext.pdf</uri></json:item><json:item><extension>zip</extension><original>false</original><mimetype>application/zip</mimetype><uri>https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/bundle.zip</uri></json:item><istex:fulltextTEI uri="https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/fulltext.tei"><teiHeader><fileDesc><titleStmt><title level="a" type="main" xml:lang="en">A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title></titleStmt><publicationStmt><authority>ISTEX</authority><availability><licence>Springer-Verlag Berlin Heidelberg</licence></availability><date when="2005">2005</date></publicationStmt><notesStmt><note type="conference" source="proceedings" scheme="https://content-type.data.istex.fr/ark:/67375/XTP-BFHXPBJJ-3">conference</note><note type="publication-type" subtype="book-series" scheme="https://publication-type.data.istex.fr/ark:/67375/JMC-0G6R5W5T-Z">book-series</note></notesStmt><sourceDesc><biblStruct><analytic><title level="a" type="main" xml:lang="en">A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title><author><persName><forename type="first">Hubert</forename><surname>Cecotti</surname></persName><email>hubert.cecotti@loria.fr</email><affiliation><orgName type="institution">LORIA/CNRS</orgName><address><street>Campus Scientifique</street><postBox>BP 239</postBox><postCode>54506</postCode><settlement>Vandoeuvre-les-Nancy cedex</settlement><country key="FR">FRANCE</country></address></affiliation></author><author><persName><forename type="first">Abdel</forename><surname>Belaïd</surname></persName><email>abdel.belaid@loria.fr</email><affiliation><orgName type="institution">LORIA/CNRS</orgName><address><street>Campus Scientifique</street><postBox>BP 239</postBox><postCode>54506</postCode><settlement>Vandoeuvre-les-Nancy cedex</settlement><country key="FR">FRANCE</country></address></affiliation></author><idno type="istex">CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9</idno><idno type="ark">ark:/67375/HCB-WMVS30DS-M</idno><idno type="DOI">10.1007/3-540-32390-2_13</idno></analytic><monogr><title level="m" type="main">Computer Recognition Systems</title><title level="m" type="sub">Proceedings of the 4th International Conference on Computer Recognition Systems CORES ’05</title><title level="m" type="part">Features, Learning and Classifiers</title><idno type="DOI">10.1007/3-540-32390-2</idno><idno type="book-id">978-3-540-32390-7</idno><idno type="ISBN">978-3-540-25054-8</idno><idno type="eISBN">978-3-540-32390-7</idno><idno type="chapter-id">Chap13</idno><idno type="part-id">Part2</idno><editor><persName><forename type="first">Marek</forename><surname>Kurzyński</surname></persName><affiliation><orgName type="department">Faculty of Electronics</orgName><orgName type="institution">Wroclaw University of Technology</orgName><address><street>Wybrzeze Wyspianskiego 27</street><postCode>50-370</postCode><settlement>Wroclaw</settlement><country key="PL">POLAND</country></address></affiliation></editor><editor><persName><forename type="first">Edward</forename><surname>Puchała</surname></persName><affiliation><orgName type="department">Faculty of Electronics</orgName><orgName type="institution">Wroclaw University of Technology</orgName><address><street>Wybrzeze Wyspianskiego 27</street><postCode>50-370</postCode><settlement>Wroclaw</settlement><country key="PL">POLAND</country></address></affiliation></editor><editor><persName><forename type="first">Michał</forename><surname>Woźniak</surname></persName><affiliation><orgName type="department">Faculty of Electronics</orgName><orgName type="institution">Wroclaw University of Technology</orgName><address><street>Wybrzeze Wyspianskiego 27</street><postCode>50-370</postCode><settlement>Wroclaw</settlement><country key="PL">POLAND</country></address></affiliation></editor><editor><persName><forename type="first">Andrzej</forename><surname>żołnierek</surname></persName><affiliation><orgName type="department">Faculty of Electronics</orgName><orgName type="institution">Wroclaw University of Technology</orgName><address><street>Wybrzeze Wyspianskiego 27</street><postCode>50-370</postCode><settlement>Wroclaw</settlement><country key="PL">POLAND</country></address></affiliation></editor><imprint><biblScope unit="vol">30</biblScope><biblScope unit="page" from="129">129</biblScope><biblScope unit="page" to="136">136</biblScope><biblScope unit="chapter-count">105</biblScope><biblScope unit="part-chapter-count">34</biblScope></imprint></monogr><series><title level="s" type="main" xml:lang="en">Advances in Soft Computing</title><editor><persName><forename type="first">Janusz</forename><surname>Kacprzyk</surname></persName><email>kacprzyk@ibspan.waw.pl</email><affiliation><orgName type="department">Systems Research Institute</orgName><orgName type="institution">Polish Academy of Sciences</orgName><address><street>ul. Newelska 6</street><postCode>01-447</postCode><settlement>Warsaw</settlement><country key="PL">POLAND</country></address></affiliation></editor><idno type="pISSN">1615-3871</idno><idno type="seriesID">4240</idno></series></biblStruct></sourceDesc></fileDesc><profileDesc><abstract xml:lang="en"><head>Abstract</head><p>We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.</p></abstract><textClass ana="subject"><keywords scheme="book-subject-collection"><list><label>SUCO11647</label><item><term>Engineering</term></item></list></keywords></textClass><textClass ana="subject"><keywords scheme="book-subject"><list><label>SCI</label><item><term type="Primary">Computer Science</term></item><label>SCI2203X</label><item><term type="Secondary" subtype="priority-1">Pattern Recognition</term></item><label>SCI21017</label><item><term type="Secondary" subtype="priority-2">Artificial Intelligence (incl. Robotics)</term></item><label>SCT11006</label><item><term type="Secondary" subtype="priority-3">Appl.Mathematics/Computational Methods of Engineering</term></item><label>SCM13003</label><item><term type="Secondary" subtype="priority-4">Applications of Mathematics</term></item><label>SCI18008</label><item><term type="Secondary" subtype="priority-5">Information Systems and Communication Service</term></item></list></keywords></textClass><langUsage><language ident="EN"></language></langUsage></profileDesc></teiHeader></istex:fulltextTEI><json:item><extension>txt</extension><original>false</original><mimetype>text/plain</mimetype><uri>https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/fulltext.txt</uri></json:item></fulltext><metadata><istex:metadataXml wicri:clean="corpus springer-ebooks not found" wicri:toSee="no header"><istex:xmlDeclaration>version="1.0" encoding="UTF-8"</istex:xmlDeclaration><istex:docType PUBLIC="-//Springer-Verlag//DTD A++ V2.4//EN" URI="http://devel.springer.de/A++/V2.4/DTD/A++V2.4.dtd" name="istex:docType"></istex:docType><istex:document><Publisher><PublisherInfo><PublisherName>Springer Berlin Heidelberg</PublisherName><PublisherLocation>Berlin, Heidelberg</PublisherLocation></PublisherInfo><Series><SeriesInfo SeriesType="Series" TocLevels="0"><SeriesID>4240</SeriesID><SeriesPrintISSN>1615-3871</SeriesPrintISSN><SeriesTitle Language="En">Advances in Soft Computing</SeriesTitle></SeriesInfo><SeriesHeader><EditorGroup><Editor AffiliationIDS="Aff1"><EditorName DisplayOrder="Western"><Prefix>Prof.</Prefix><GivenName>Janusz</GivenName><FamilyName>Kacprzyk</FamilyName></EditorName><Contact><Email>kacprzyk@ibspan.waw.pl</Email></Contact></Editor><Affiliation ID="Aff1"><OrgDivision>Systems Research Institute</OrgDivision><OrgName>Polish Academy of Sciences</OrgName><OrgAddress><Street>ul. Newelska 6</Street><Postcode>01-447</Postcode><City>Warsaw</City><Country>Poland</Country></OrgAddress></Affiliation></EditorGroup></SeriesHeader><Book Language="En"><BookInfo BookProductType="Proceedings" ContainsESM="No" Language="En" MediaType="eBook" NumberingStyle="Unnumbered" TocLevels="0"><BookID>978-3-540-32390-7</BookID><BookTitle>Computer Recognition Systems</BookTitle><BookSubTitle>Proceedings of the 4th International Conference on Computer Recognition Systems CORES ’05</BookSubTitle><BookVolumeNumber>30</BookVolumeNumber><BookSequenceNumber>30</BookSequenceNumber><BookDOI>10.1007/3-540-32390-2</BookDOI><BookTitleID>115303</BookTitleID><BookPrintISBN>978-3-540-25054-8</BookPrintISBN><BookElectronicISBN>978-3-540-32390-7</BookElectronicISBN><BookChapterCount>105</BookChapterCount><BookCopyright><CopyrightHolderName>Springer-Verlag Berlin Heidelberg</CopyrightHolderName><CopyrightYear>2005</CopyrightYear></BookCopyright><BookSubjectGroup><BookSubject Code="SCI" Type="Primary">Computer Science</BookSubject><BookSubject Code="SCI2203X" Priority="1" Type="Secondary">Pattern Recognition</BookSubject><BookSubject Code="SCI21017" Priority="2" Type="Secondary">Artificial Intelligence (incl. Robotics)</BookSubject><BookSubject Code="SCT11006" Priority="3" Type="Secondary">Appl.Mathematics/Computational Methods of Engineering</BookSubject><BookSubject Code="SCM13003" Priority="4" Type="Secondary">Applications of Mathematics</BookSubject><BookSubject Code="SCI18008" Priority="5" Type="Secondary">Information Systems and Communication Service</BookSubject><SubjectCollection Code="SUCO11647">Engineering</SubjectCollection></BookSubjectGroup><BookContext><SeriesID>4240</SeriesID></BookContext></BookInfo><BookHeader><EditorGroup><Editor AffiliationIDS="Aff2"><EditorName DisplayOrder="Western"><Prefix>Prof.</Prefix><GivenName>Marek</GivenName><FamilyName>Kurzyński</FamilyName></EditorName></Editor><Editor AffiliationIDS="Aff2"><EditorName DisplayOrder="Western"><Prefix>Dr.</Prefix><GivenName>Edward</GivenName><FamilyName>Puchała</FamilyName></EditorName></Editor><Editor AffiliationIDS="Aff2"><EditorName DisplayOrder="Western"><Prefix>Dr.</Prefix><GivenName>Michał</GivenName><FamilyName>Woźniak</FamilyName></EditorName></Editor><Editor AffiliationIDS="Aff2"><EditorName DisplayOrder="Western"><Prefix>Dr.</Prefix><GivenName>Andrzej</GivenName><FamilyName>żołnierek</FamilyName></EditorName></Editor><Affiliation ID="Aff2"><OrgDivision>Faculty of Electronics</OrgDivision><OrgName>Wroclaw University of Technology</OrgName><OrgAddress><Street>Wybrzeze Wyspianskiego 27</Street><Postcode>50-370</Postcode><City>Wroclaw</City><Country>Poland</Country></OrgAddress></Affiliation></EditorGroup></BookHeader><Part ID="Part2"><PartInfo TocLevels="0"><PartID>2</PartID><PartNumber>Part II</PartNumber><PartSequenceNumber>2</PartSequenceNumber><PartTitle>Features, Learning and Classifiers</PartTitle><PartChapterCount>34</PartChapterCount><PartContext><SeriesID>4240</SeriesID><BookID>3-540-32390-2</BookID><BookTitle>Computer Recognition Systems</BookTitle></PartContext></PartInfo><Chapter ID="Chap13" Language="En"><ChapterInfo ChapterType="OriginalPaper" ContainsESM="No" Language="En" NumberingStyle="Unnumbered" TocLevels="0"><ChapterID>13</ChapterID><ChapterDOI>10.1007/3-540-32390-2_13</ChapterDOI><ChapterSequenceNumber>13</ChapterSequenceNumber><ChapterTitle Language="En">A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</ChapterTitle><ChapterFirstPage>129</ChapterFirstPage><ChapterLastPage>136</ChapterLastPage><ChapterCopyright><CopyrightHolderName>Springer-Verlag Berlin Heidelberg</CopyrightHolderName><CopyrightYear>2005</CopyrightYear></ChapterCopyright><ChapterGrants Type="Regular"><MetadataGrant Grant="OpenAccess"></MetadataGrant><AbstractGrant Grant="OpenAccess"></AbstractGrant><BodyPDFGrant Grant="Restricted"></BodyPDFGrant><BodyHTMLGrant Grant="Restricted"></BodyHTMLGrant><BibliographyGrant Grant="Restricted"></BibliographyGrant><ESMGrant Grant="Restricted"></ESMGrant></ChapterGrants><ChapterContext><SeriesID>4240</SeriesID><PartID>2</PartID><BookID>3-540-32390-2</BookID><BookTitle>Computer Recognition Systems</BookTitle></ChapterContext></ChapterInfo><ChapterHeader><AuthorGroup><Author AffiliationIDS="Aff3"><AuthorName DisplayOrder="Western"><GivenName>Hubert</GivenName><FamilyName>Cecotti</FamilyName></AuthorName><Contact><Email>hubert.cecotti@loria.fr</Email></Contact></Author><Author AffiliationIDS="Aff3"><AuthorName DisplayOrder="Western"><GivenName>Abdel</GivenName><FamilyName>Belaïd</FamilyName></AuthorName><Contact><Email>abdel.belaid@loria.fr</Email></Contact></Author><Affiliation ID="Aff3"><OrgName>LORIA/CNRS</OrgName><OrgAddress><Street>Campus Scientifique</Street><Postbox>BP 239</Postbox><Postcode>54506</Postcode><City>Vandoeuvre-les-Nancy cedex</City><Country>France</Country></OrgAddress></Affiliation></AuthorGroup><Abstract ID="Abs1" Language="En"><Heading>Abstract</Heading><Para TextBreak="No">We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.</Para></Abstract></ChapterHeader><NoBody></NoBody></Chapter></Part></Book></Series></Publisher></istex:document></istex:metadataXml><mods version="3.6"><titleInfo lang="en"><title>A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title></titleInfo><titleInfo type="alternative" contentType="CDATA"><title>A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology</title></titleInfo><name type="personal"><namePart type="given">Hubert</namePart><namePart type="family">Cecotti</namePart><affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</affiliation><affiliation>E-mail: hubert.cecotti@loria.fr</affiliation><role><roleTerm type="text">author</roleTerm></role></name><name type="personal"><namePart type="given">Abdel</namePart><namePart type="family">Belaïd</namePart><affiliation>LORIA/CNRS, Campus Scientifique, BP 239, 54506, Vandoeuvre-les-Nancy cedex, France</affiliation><affiliation>E-mail: abdel.belaid@loria.fr</affiliation><role><roleTerm type="text">author</roleTerm></role></name><typeOfResource>text</typeOfResource><genre displayLabel="OriginalPaper" authority="ISTEX" authorityURI="https://content-type.data.istex.fr" type="conference" valueURI="https://content-type.data.istex.fr/ark:/67375/XTP-BFHXPBJJ-3">conference</genre><originInfo><publisher>Springer Berlin Heidelberg</publisher><place><placeTerm type="text">Berlin, Heidelberg</placeTerm></place><dateIssued encoding="w3cdtf">2005</dateIssued><copyrightDate encoding="w3cdtf">2005</copyrightDate></originInfo><language><languageTerm type="code" authority="rfc3066">en</languageTerm><languageTerm type="code" authority="iso639-2b">eng</languageTerm></language><abstract lang="en">Abstract: We presented a model combining supervised learning for the classification stage and a specific not-supervised model for the test stage. It allows changing the topology of the supervised part for improving the reject quality. Within the framework of the recognition of disturbed forms, our strategy was based on the adaptability of the classifier instead of, for example, making it learn more samples by artificially creating all the possible deformations, our approach consisted in defining the type of problematical transformation, by changing the localities observed by the principal classifier. The system makes it possible indeed to refine the results till it makes possible to raise ambiguities for certain confusions but there remains still depend on the effectiveness of the initial rejection and the assumptions concerning the choice of the model of class. On the level of the prospects, if the type of deformation is perfectly known: cut, rotation, problem of shift, it would be possible to introduce new properties inside the self-organizing map to speed up its convergence.</abstract><relatedItem type="host"><titleInfo><title>Computer Recognition Systems</title><subTitle>Proceedings of the 4th International Conference on Computer Recognition Systems CORES ’05</subTitle></titleInfo><name type="personal"><namePart type="termsOfAddress">Prof.</namePart><namePart type="given">Marek</namePart><namePart type="family">Kurzyński</namePart><affiliation>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="termsOfAddress">Dr.</namePart><namePart type="given">Edward</namePart><namePart type="family">Puchała</namePart><affiliation>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="termsOfAddress">Dr.</namePart><namePart type="given">Michał</namePart><namePart type="family">Woźniak</namePart><affiliation>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><name type="personal"><namePart type="termsOfAddress">Dr.</namePart><namePart type="given">Andrzej</namePart><namePart type="family">żołnierek</namePart><affiliation>Faculty of Electronics, Wroclaw University of Technology, Wybrzeze Wyspianskiego 27, 50-370, Wroclaw, Poland</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><genre type="book-series" authority="ISTEX" authorityURI="https://publication-type.data.istex.fr" valueURI="https://publication-type.data.istex.fr/ark:/67375/JMC-0G6R5W5T-Z">book-series</genre><originInfo><publisher>Springer</publisher><copyrightDate encoding="w3cdtf">2005</copyrightDate><issuance>monographic</issuance></originInfo><subject><genre>Book-Subject-Collection</genre><topic authority="SpringerSubjectCodes" authorityURI="SUCO11647">Engineering</topic></subject><subject><genre>Book-Subject-Group</genre><topic authority="SpringerSubjectCodes" authorityURI="SCI">Computer Science</topic><topic authority="SpringerSubjectCodes" authorityURI="SCI2203X">Pattern Recognition</topic><topic authority="SpringerSubjectCodes" authorityURI="SCI21017">Artificial Intelligence (incl. Robotics)</topic><topic authority="SpringerSubjectCodes" authorityURI="SCT11006">Appl.Mathematics/Computational Methods of Engineering</topic><topic authority="SpringerSubjectCodes" authorityURI="SCM13003">Applications of Mathematics</topic><topic authority="SpringerSubjectCodes" authorityURI="SCI18008">Information Systems and Communication Service</topic></subject><identifier type="DOI">10.1007/3-540-32390-2</identifier><identifier type="ISBN">978-3-540-25054-8</identifier><identifier type="eISBN">978-3-540-32390-7</identifier><identifier type="ISSN">1615-3871</identifier><identifier type="BookTitleID">115303</identifier><identifier type="BookID">978-3-540-32390-7</identifier><identifier type="BookChapterCount">105</identifier><identifier type="BookVolumeNumber">30</identifier><identifier type="BookSequenceNumber">30</identifier><identifier type="PartChapterCount">34</identifier><part><date>2005</date><detail type="part"><title>Part II: Features, Learning and Classifiers</title></detail><detail type="volume"><number>30</number><caption>vol.</caption></detail><extent unit="pages"><start>129</start><end>136</end></extent></part><recordInfo><recordOrigin>Springer-Verlag Berlin Heidelberg, 2005</recordOrigin></recordInfo></relatedItem><relatedItem type="series"><titleInfo><title>Advances in Soft Computing</title></titleInfo><name type="personal"><namePart type="termsOfAddress">Prof.</namePart><namePart type="given">Janusz</namePart><namePart type="family">Kacprzyk</namePart><affiliation>Systems Research Institute, Polish Academy of Sciences, ul. Newelska 6, 01-447, Warsaw, Poland</affiliation><affiliation>E-mail: kacprzyk@ibspan.waw.pl</affiliation><role><roleTerm type="text">editor</roleTerm></role></name><originInfo><publisher>Springer</publisher><copyrightDate encoding="w3cdtf">2005</copyrightDate><issuance>serial</issuance></originInfo><identifier type="ISSN">1615-3871</identifier><identifier type="SeriesID">4240</identifier><part><detail type="volume"><number>Part II</number><caption>vol.</caption></detail></part><recordInfo><recordOrigin>Springer-Verlag Berlin Heidelberg, 2005</recordOrigin></recordInfo></relatedItem><identifier type="istex">CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9</identifier><identifier type="ark">ark:/67375/HCB-WMVS30DS-M</identifier><identifier type="DOI">10.1007/3-540-32390-2_13</identifier><identifier type="ChapterID">13</identifier><identifier type="ChapterID">Chap13</identifier><accessCondition type="use and reproduction" contentType="copyright">Springer-Verlag Berlin Heidelberg, 2005</accessCondition><recordInfo><recordContentSource authority="ISTEX" authorityURI="https://loaded-corpus.data.istex.fr" valueURI="https://loaded-corpus.data.istex.fr/ark:/67375/XBH-RLRX46XW-4">springer</recordContentSource><recordOrigin>Springer-Verlag Berlin Heidelberg, 2005</recordOrigin></recordInfo></mods><json:item><extension>json</extension><original>false</original><mimetype>application/json</mimetype><uri>https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/record.json</uri></json:item></metadata><annexes><json:item><extension>txt</extension><original>true</original><mimetype>text/plain</mimetype><uri>https://api.istex.fr/ark:/67375/HCB-WMVS30DS-M/annexes.txt</uri></json:item></annexes></istex></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Lorraine/explor/InforLorV4/Data/Istex/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 003023 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/Istex/Corpus/biblio.hfd -nk 003023 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Lorraine
|area= InforLorV4
|flux= Istex
|étape= Corpus
|type= RBID
|clé= ISTEX:CB2F12B757C4CC408E2EC8EA049A7B05CEDE71A9
|texte= A New Rejection Strategy for Convolutional Neural Network by Adaptive Topology
}}
| This area was generated with Dilib version V0.6.33. |  |