Matrix Completion for Weakly-Supervised Multi-Label Image Classification

In the last few years, image classification has become an incredibly active research topic, with widespread applications. Most methods for visual recognition are fully supervised, as they make use of bounding boxes or pixelwise segmentations to locate objects of interest. However, this type of manua...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on pattern analysis and machine intelligence 2015-01, Vol.37 (1), p.121-135
Hauptverfasser: Cabral, Ricardo, De la Torre, Fernando, Costeira, Joao Paulo, Bernardino, Alexandre
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 135
container_issue 1
container_start_page 121
container_title IEEE transactions on pattern analysis and machine intelligence
container_volume 37
creator Cabral, Ricardo
De la Torre, Fernando
Costeira, Joao Paulo
Bernardino, Alexandre
description In the last few years, image classification has become an incredibly active research topic, with widespread applications. Most methods for visual recognition are fully supervised, as they make use of bounding boxes or pixelwise segmentations to locate objects of interest. However, this type of manual labeling is time consuming, error prone and it has been shown that manual segmentations are not necessarily the optimal spatial enclosure for object classifiers. This paper proposes a weakly-supervised system for multi-label image classification. In this setting, training images are annotated with a set of keywords describing their contents, but the visual concepts are not explicitly segmented in the images. We formulate the weakly-supervised image classification as a low-rank matrix completion problem. Compared to previous work, our proposed framework has three advantages: (1) Unlike existing solutions based on multiple-instance learning methods, our model is convex. We propose two alternative algorithms for matrix completion specifically tailored to visual data, and prove their convergence. (2) Unlike existing discriminative methods, our algorithm is robust to labeling errors, background noise and partial occlusions. (3) Our method can potentially be used for semantic segmentation. Experimental validation on several data sets shows that our method outperforms state-of-the-art classification algorithms, while effectively capturing each class appearance.
doi_str_mv 10.1109/TPAMI.2014.2343234
format Article
fullrecord <record><control><sourceid>proquest_RIE</sourceid><recordid>TN_cdi_ieee_primary_6866218</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>6866218</ieee_id><sourcerecordid>1711534633</sourcerecordid><originalsourceid>FETCH-LOGICAL-c384t-8b468cf0cb3dfa91ed878e3667467585360c0184daad714577f05278522593e63</originalsourceid><addsrcrecordid>eNqNkU1LAzEQhoMoWj_-gIIsePGyNclsPvYoxY9Ci4KKx5Duzkp0t1uTXbH_3tRWD548DHOY5x0meQg5ZnTIGM0vHu8vp-Mhpywbcsgg1hYZsBzyFATk22RAmeSp1lzvkf0QXmkkBYVdssdlJDiDAbmd2s67z2TUNosaO9fOk6r1yTPat3qZPvQL9B8uYJlM-7pz6cTOsE7GjX3BZFTbEFzlCruKHZKdytYBjzb9gDxdXz2ObtPJ3c14dDlJC9BZl-pZJnVR0WIGZWVzhqVWGkFKlUkltABJC8p0VlpbqniuUhUVXGnBucgBJRyQ8_XehW_fewydaVwosK7tHNs-GKYkp0px9R-UMQGZBIjo2R_0te39PD7EMBk_i0mdiUjxNVX4NgSPlVl411i_NIyalRLzrcSslJiNkhg63azuZw2Wv5EfBxE4WQMOEX_HUkvJmYYvLmWNAA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>1663516845</pqid></control><display><type>article</type><title>Matrix Completion for Weakly-Supervised Multi-Label Image Classification</title><source>IEEE Electronic Library (IEL)</source><creator>Cabral, Ricardo ; De la Torre, Fernando ; Costeira, Joao Paulo ; Bernardino, Alexandre</creator><creatorcontrib>Cabral, Ricardo ; De la Torre, Fernando ; Costeira, Joao Paulo ; Bernardino, Alexandre</creatorcontrib><description>In the last few years, image classification has become an incredibly active research topic, with widespread applications. Most methods for visual recognition are fully supervised, as they make use of bounding boxes or pixelwise segmentations to locate objects of interest. However, this type of manual labeling is time consuming, error prone and it has been shown that manual segmentations are not necessarily the optimal spatial enclosure for object classifiers. This paper proposes a weakly-supervised system for multi-label image classification. In this setting, training images are annotated with a set of keywords describing their contents, but the visual concepts are not explicitly segmented in the images. We formulate the weakly-supervised image classification as a low-rank matrix completion problem. Compared to previous work, our proposed framework has three advantages: (1) Unlike existing solutions based on multiple-instance learning methods, our model is convex. We propose two alternative algorithms for matrix completion specifically tailored to visual data, and prove their convergence. (2) Unlike existing discriminative methods, our algorithm is robust to labeling errors, background noise and partial occlusions. (3) Our method can potentially be used for semantic segmentation. Experimental validation on several data sets shows that our method outperforms state-of-the-art classification algorithms, while effectively capturing each class appearance.</description><identifier>ISSN: 0162-8828</identifier><identifier>EISSN: 1939-3539</identifier><identifier>EISSN: 2160-9292</identifier><identifier>DOI: 10.1109/TPAMI.2014.2343234</identifier><identifier>PMID: 26353213</identifier><identifier>CODEN: ITPIDJ</identifier><language>eng</language><publisher>United States: IEEE</publisher><subject>Algorithms ; Classification ; Histograms ; Image classification ; Image segmentation ; Manuals ; Marking ; Mathematical models ; Methods ; Minimization ; Pattern analysis ; Segmentation ; Semantics ; Training ; Vectors ; Visual</subject><ispartof>IEEE transactions on pattern analysis and machine intelligence, 2015-01, Vol.37 (1), p.121-135</ispartof><rights>Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2015</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c384t-8b468cf0cb3dfa91ed878e3667467585360c0184daad714577f05278522593e63</citedby><cites>FETCH-LOGICAL-c384t-8b468cf0cb3dfa91ed878e3667467585360c0184daad714577f05278522593e63</cites><orcidid>0000-0002-4919-8711</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/6866218$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,796,27924,27925,54758</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/6866218$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/26353213$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Cabral, Ricardo</creatorcontrib><creatorcontrib>De la Torre, Fernando</creatorcontrib><creatorcontrib>Costeira, Joao Paulo</creatorcontrib><creatorcontrib>Bernardino, Alexandre</creatorcontrib><title>Matrix Completion for Weakly-Supervised Multi-Label Image Classification</title><title>IEEE transactions on pattern analysis and machine intelligence</title><addtitle>TPAMI</addtitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><description>In the last few years, image classification has become an incredibly active research topic, with widespread applications. Most methods for visual recognition are fully supervised, as they make use of bounding boxes or pixelwise segmentations to locate objects of interest. However, this type of manual labeling is time consuming, error prone and it has been shown that manual segmentations are not necessarily the optimal spatial enclosure for object classifiers. This paper proposes a weakly-supervised system for multi-label image classification. In this setting, training images are annotated with a set of keywords describing their contents, but the visual concepts are not explicitly segmented in the images. We formulate the weakly-supervised image classification as a low-rank matrix completion problem. Compared to previous work, our proposed framework has three advantages: (1) Unlike existing solutions based on multiple-instance learning methods, our model is convex. We propose two alternative algorithms for matrix completion specifically tailored to visual data, and prove their convergence. (2) Unlike existing discriminative methods, our algorithm is robust to labeling errors, background noise and partial occlusions. (3) Our method can potentially be used for semantic segmentation. Experimental validation on several data sets shows that our method outperforms state-of-the-art classification algorithms, while effectively capturing each class appearance.</description><subject>Algorithms</subject><subject>Classification</subject><subject>Histograms</subject><subject>Image classification</subject><subject>Image segmentation</subject><subject>Manuals</subject><subject>Marking</subject><subject>Mathematical models</subject><subject>Methods</subject><subject>Minimization</subject><subject>Pattern analysis</subject><subject>Segmentation</subject><subject>Semantics</subject><subject>Training</subject><subject>Vectors</subject><subject>Visual</subject><issn>0162-8828</issn><issn>1939-3539</issn><issn>2160-9292</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2015</creationdate><recordtype>article</recordtype><sourceid>RIE</sourceid><recordid>eNqNkU1LAzEQhoMoWj_-gIIsePGyNclsPvYoxY9Ci4KKx5Duzkp0t1uTXbH_3tRWD548DHOY5x0meQg5ZnTIGM0vHu8vp-Mhpywbcsgg1hYZsBzyFATk22RAmeSp1lzvkf0QXmkkBYVdssdlJDiDAbmd2s67z2TUNosaO9fOk6r1yTPat3qZPvQL9B8uYJlM-7pz6cTOsE7GjX3BZFTbEFzlCruKHZKdytYBjzb9gDxdXz2ObtPJ3c14dDlJC9BZl-pZJnVR0WIGZWVzhqVWGkFKlUkltABJC8p0VlpbqniuUhUVXGnBucgBJRyQ8_XehW_fewydaVwosK7tHNs-GKYkp0px9R-UMQGZBIjo2R_0te39PD7EMBk_i0mdiUjxNVX4NgSPlVl411i_NIyalRLzrcSslJiNkhg63azuZw2Wv5EfBxE4WQMOEX_HUkvJmYYvLmWNAA</recordid><startdate>20150101</startdate><enddate>20150101</enddate><creator>Cabral, Ricardo</creator><creator>De la Torre, Fernando</creator><creator>Costeira, Joao Paulo</creator><creator>Bernardino, Alexandre</creator><general>IEEE</general><general>The Institute of Electrical and Electronics Engineers, Inc. (IEEE)</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7SC</scope><scope>7SP</scope><scope>8FD</scope><scope>JQ2</scope><scope>L7M</scope><scope>L~C</scope><scope>L~D</scope><scope>7X8</scope><scope>F28</scope><scope>FR3</scope><orcidid>https://orcid.org/0000-0002-4919-8711</orcidid></search><sort><creationdate>20150101</creationdate><title>Matrix Completion for Weakly-Supervised Multi-Label Image Classification</title><author>Cabral, Ricardo ; De la Torre, Fernando ; Costeira, Joao Paulo ; Bernardino, Alexandre</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c384t-8b468cf0cb3dfa91ed878e3667467585360c0184daad714577f05278522593e63</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2015</creationdate><topic>Algorithms</topic><topic>Classification</topic><topic>Histograms</topic><topic>Image classification</topic><topic>Image segmentation</topic><topic>Manuals</topic><topic>Marking</topic><topic>Mathematical models</topic><topic>Methods</topic><topic>Minimization</topic><topic>Pattern analysis</topic><topic>Segmentation</topic><topic>Semantics</topic><topic>Training</topic><topic>Vectors</topic><topic>Visual</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Cabral, Ricardo</creatorcontrib><creatorcontrib>De la Torre, Fernando</creatorcontrib><creatorcontrib>Costeira, Joao Paulo</creatorcontrib><creatorcontrib>Bernardino, Alexandre</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005-present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998-Present</collection><collection>IEEE Electronic Library (IEL)</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Computer and Information Systems Abstracts</collection><collection>Electronics &amp; Communications Abstracts</collection><collection>Technology Research Database</collection><collection>ProQuest Computer Science Collection</collection><collection>Advanced Technologies Database with Aerospace</collection><collection>Computer and Information Systems Abstracts – Academic</collection><collection>Computer and Information Systems Abstracts Professional</collection><collection>MEDLINE - Academic</collection><collection>ANTE: Abstracts in New Technology &amp; Engineering</collection><collection>Engineering Research Database</collection><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Cabral, Ricardo</au><au>De la Torre, Fernando</au><au>Costeira, Joao Paulo</au><au>Bernardino, Alexandre</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Matrix Completion for Weakly-Supervised Multi-Label Image Classification</atitle><jtitle>IEEE transactions on pattern analysis and machine intelligence</jtitle><stitle>TPAMI</stitle><addtitle>IEEE Trans Pattern Anal Mach Intell</addtitle><date>2015-01-01</date><risdate>2015</risdate><volume>37</volume><issue>1</issue><spage>121</spage><epage>135</epage><pages>121-135</pages><issn>0162-8828</issn><eissn>1939-3539</eissn><eissn>2160-9292</eissn><coden>ITPIDJ</coden><abstract>In the last few years, image classification has become an incredibly active research topic, with widespread applications. Most methods for visual recognition are fully supervised, as they make use of bounding boxes or pixelwise segmentations to locate objects of interest. However, this type of manual labeling is time consuming, error prone and it has been shown that manual segmentations are not necessarily the optimal spatial enclosure for object classifiers. This paper proposes a weakly-supervised system for multi-label image classification. In this setting, training images are annotated with a set of keywords describing their contents, but the visual concepts are not explicitly segmented in the images. We formulate the weakly-supervised image classification as a low-rank matrix completion problem. Compared to previous work, our proposed framework has three advantages: (1) Unlike existing solutions based on multiple-instance learning methods, our model is convex. We propose two alternative algorithms for matrix completion specifically tailored to visual data, and prove their convergence. (2) Unlike existing discriminative methods, our algorithm is robust to labeling errors, background noise and partial occlusions. (3) Our method can potentially be used for semantic segmentation. Experimental validation on several data sets shows that our method outperforms state-of-the-art classification algorithms, while effectively capturing each class appearance.</abstract><cop>United States</cop><pub>IEEE</pub><pmid>26353213</pmid><doi>10.1109/TPAMI.2014.2343234</doi><tpages>15</tpages><orcidid>https://orcid.org/0000-0002-4919-8711</orcidid></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 0162-8828
ispartof IEEE transactions on pattern analysis and machine intelligence, 2015-01, Vol.37 (1), p.121-135
issn 0162-8828
1939-3539
2160-9292
language eng
recordid cdi_ieee_primary_6866218
source IEEE Electronic Library (IEL)
subjects Algorithms
Classification
Histograms
Image classification
Image segmentation
Manuals
Marking
Mathematical models
Methods
Minimization
Pattern analysis
Segmentation
Semantics
Training
Vectors
Visual
title Matrix Completion for Weakly-Supervised Multi-Label Image Classification
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-25T02%3A14%3A31IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_RIE&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Matrix%20Completion%20for%20Weakly-Supervised%20Multi-Label%20Image%20Classification&rft.jtitle=IEEE%20transactions%20on%20pattern%20analysis%20and%20machine%20intelligence&rft.au=Cabral,%20Ricardo&rft.date=2015-01-01&rft.volume=37&rft.issue=1&rft.spage=121&rft.epage=135&rft.pages=121-135&rft.issn=0162-8828&rft.eissn=1939-3539&rft.coden=ITPIDJ&rft_id=info:doi/10.1109/TPAMI.2014.2343234&rft_dat=%3Cproquest_RIE%3E1711534633%3C/proquest_RIE%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=1663516845&rft_id=info:pmid/26353213&rft_ieee_id=6866218&rfr_iscdi=true