Probability Models for Open Set Recognition.
Identifieur interne : 000005 ( PubMed/Corpus ); précédent : 000004; suivant : 000006Probability Models for Open Set Recognition.
Auteurs : Walter J. Scheirer ; Lalit P. Jain ; Terrance E. BoultSource :
- IEEE transactions on pattern analysis and machine intelligence [ 1939-3539 ] ; 2014.
Abstract
Real-world tasks in computer vision often touch upon open set recognition: multi-class recognition with incomplete knowledge of the world and many unknown inputs. Recent work on this problem has proposed a model incorporating an open space risk term to account for the space beyond the reasonable support of known classes. This paper extends the general idea of open space risk limiting classification to accommodate non-linear classifiers in a multiclass setting. We introduce a new open set recognition model called compact abating probability (CAP), where the probability of class membership decreases in value (abates) as points move from known data toward open space. We show that CAP models improve open set recognition for multiple algorithms. Leveraging the CAP formulation, we go on to describe the novel Weibull-calibrated SVM (W-SVM) algorithm, which combines the useful properties of statistical extreme value theory for score calibration with one-class and binary support vector machines. Our experiments show that the W-SVM is significantly better for open set object detection and OCR problems when compared to the state-of-the-art for the same tasks.
DOI: 10.1109/TPAMI.2014.2321392
PubMed: 26353070
Links to Exploration step
pubmed:26353070Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Probability Models for Open Set Recognition.</title>
<author><name sortKey="Scheirer, Walter J" sort="Scheirer, Walter J" uniqKey="Scheirer W" first="Walter J" last="Scheirer">Walter J. Scheirer</name>
</author>
<author><name sortKey="Jain, Lalit P" sort="Jain, Lalit P" uniqKey="Jain L" first="Lalit P" last="Jain">Lalit P. Jain</name>
</author>
<author><name sortKey="Boult, Terrance E" sort="Boult, Terrance E" uniqKey="Boult T" first="Terrance E" last="Boult">Terrance E. Boult</name>
</author>
</titleStmt>
<publicationStmt><idno type="wicri:source">PubMed</idno>
<date when="2014">2014</date>
<idno type="doi">10.1109/TPAMI.2014.2321392</idno>
<idno type="RBID">pubmed:26353070</idno>
<idno type="pmid">26353070</idno>
<idno type="wicri:Area/PubMed/Corpus">000005</idno>
</publicationStmt>
<sourceDesc><biblStruct><analytic><title xml:lang="en">Probability Models for Open Set Recognition.</title>
<author><name sortKey="Scheirer, Walter J" sort="Scheirer, Walter J" uniqKey="Scheirer W" first="Walter J" last="Scheirer">Walter J. Scheirer</name>
</author>
<author><name sortKey="Jain, Lalit P" sort="Jain, Lalit P" uniqKey="Jain L" first="Lalit P" last="Jain">Lalit P. Jain</name>
</author>
<author><name sortKey="Boult, Terrance E" sort="Boult, Terrance E" uniqKey="Boult T" first="Terrance E" last="Boult">Terrance E. Boult</name>
</author>
</analytic>
<series><title level="j">IEEE transactions on pattern analysis and machine intelligence</title>
<idno type="eISSN">1939-3539</idno>
<imprint><date when="2014" type="published">2014</date>
</imprint>
</series>
</biblStruct>
</sourceDesc>
</fileDesc>
<profileDesc><textClass></textClass>
</profileDesc>
</teiHeader>
<front><div type="abstract" xml:lang="en">Real-world tasks in computer vision often touch upon open set recognition: multi-class recognition with incomplete knowledge of the world and many unknown inputs. Recent work on this problem has proposed a model incorporating an open space risk term to account for the space beyond the reasonable support of known classes. This paper extends the general idea of open space risk limiting classification to accommodate non-linear classifiers in a multiclass setting. We introduce a new open set recognition model called compact abating probability (CAP), where the probability of class membership decreases in value (abates) as points move from known data toward open space. We show that CAP models improve open set recognition for multiple algorithms. Leveraging the CAP formulation, we go on to describe the novel Weibull-calibrated SVM (W-SVM) algorithm, which combines the useful properties of statistical extreme value theory for score calibration with one-class and binary support vector machines. Our experiments show that the W-SVM is significantly better for open set object detection and OCR problems when compared to the state-of-the-art for the same tasks.</div>
</front>
</TEI>
<pubmed><MedlineCitation Owner="NLM" Status="PubMed-not-MEDLINE"><PMID Version="1">26353070</PMID>
<DateCreated><Year>2015</Year>
<Month>09</Month>
<Day>10</Day>
</DateCreated>
<DateCompleted><Year>2015</Year>
<Month>11</Month>
<Day>25</Day>
</DateCompleted>
<Article PubModel="Print"><Journal><ISSN IssnType="Electronic">1939-3539</ISSN>
<JournalIssue CitedMedium="Internet"><Volume>36</Volume>
<Issue>11</Issue>
<PubDate><Year>2014</Year>
<Month>Nov</Month>
</PubDate>
</JournalIssue>
<Title>IEEE transactions on pattern analysis and machine intelligence</Title>
<ISOAbbreviation>IEEE Trans Pattern Anal Mach Intell</ISOAbbreviation>
</Journal>
<ArticleTitle>Probability Models for Open Set Recognition.</ArticleTitle>
<Pagination><MedlinePgn>2317-24</MedlinePgn>
</Pagination>
<ELocationID EIdType="doi" ValidYN="Y">10.1109/TPAMI.2014.2321392</ELocationID>
<Abstract><AbstractText>Real-world tasks in computer vision often touch upon open set recognition: multi-class recognition with incomplete knowledge of the world and many unknown inputs. Recent work on this problem has proposed a model incorporating an open space risk term to account for the space beyond the reasonable support of known classes. This paper extends the general idea of open space risk limiting classification to accommodate non-linear classifiers in a multiclass setting. We introduce a new open set recognition model called compact abating probability (CAP), where the probability of class membership decreases in value (abates) as points move from known data toward open space. We show that CAP models improve open set recognition for multiple algorithms. Leveraging the CAP formulation, we go on to describe the novel Weibull-calibrated SVM (W-SVM) algorithm, which combines the useful properties of statistical extreme value theory for score calibration with one-class and binary support vector machines. Our experiments show that the W-SVM is significantly better for open set object detection and OCR problems when compared to the state-of-the-art for the same tasks.</AbstractText>
</Abstract>
<AuthorList CompleteYN="Y"><Author ValidYN="Y"><LastName>Scheirer</LastName>
<ForeName>Walter J</ForeName>
<Initials>WJ</Initials>
</Author>
<Author ValidYN="Y"><LastName>Jain</LastName>
<ForeName>Lalit P</ForeName>
<Initials>LP</Initials>
</Author>
<Author ValidYN="Y"><LastName>Boult</LastName>
<ForeName>Terrance E</ForeName>
<Initials>TE</Initials>
</Author>
</AuthorList>
<Language>eng</Language>
<PublicationTypeList><PublicationType UI="D016428">Journal Article</PublicationType>
<PublicationType UI="D013485">Research Support, Non-U.S. Gov't</PublicationType>
<PublicationType UI="D013486">Research Support, U.S. Gov't, Non-P.H.S.</PublicationType>
</PublicationTypeList>
</Article>
<MedlineJournalInfo><Country>United States</Country>
<MedlineTA>IEEE Trans Pattern Anal Mach Intell</MedlineTA>
<NlmUniqueID>9885960</NlmUniqueID>
<ISSNLinking>0098-5589</ISSNLinking>
</MedlineJournalInfo>
</MedlineCitation>
<PubmedData><History><PubMedPubDate PubStatus="entrez"><Year>2015</Year>
<Month>9</Month>
<Day>10</Day>
<Hour>6</Hour>
<Minute>0</Minute>
</PubMedPubDate>
<PubMedPubDate PubStatus="pubmed"><Year>2015</Year>
<Month>9</Month>
<Day>10</Day>
<Hour>6</Hour>
<Minute>0</Minute>
</PubMedPubDate>
<PubMedPubDate PubStatus="medline"><Year>2015</Year>
<Month>9</Month>
<Day>10</Day>
<Hour>6</Hour>
<Minute>1</Minute>
</PubMedPubDate>
</History>
<PublicationStatus>ppublish</PublicationStatus>
<ArticleIdList><ArticleId IdType="doi">10.1109/TPAMI.2014.2321392</ArticleId>
<ArticleId IdType="pubmed">26353070</ArticleId>
</ArticleIdList>
</PubmedData>
</pubmed>
</record>
Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Ticri/CIDE/explor/OcrV1/Data/PubMed/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000005 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd -nk 000005 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien |wiki= Ticri/CIDE |area= OcrV1 |flux= PubMed |étape= Corpus |type= RBID |clé= pubmed:26353070 |texte= Probability Models for Open Set Recognition. }}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/RBID.i -Sk "pubmed:26353070" \ | HfdSelect -Kh $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd \ | NlmPubMed2Wicri -a OcrV1
This area was generated with Dilib version V0.6.32. |