A spectral method for context based disambiguation of image annotations
In this work we employ contextual information to improve the quality of image labellings provided by an existing automatic image annotation algorithm in a weakly supervised setting, where each training image is labelled but it is not known which part of the image its labels are referring to. We reca...
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Tagungsbericht |
| Sprache: | eng |
| Schlagworte: | |
| Online-Zugang: | Volltext bestellen |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| container_end_page | 792 |
|---|---|
| container_issue | |
| container_start_page | 789 |
| container_title | |
| container_volume | |
| creator | Semenovich, D. Sowmya, A. |
| description | In this work we employ contextual information to improve the quality of image labellings provided by an existing automatic image annotation algorithm in a weakly supervised setting, where each training image is labelled but it is not known which part of the image its labels are referring to. We recast the problem into that of constructing a graph which encodes pairwise consistency of candidate annotations and observe that mutually consistent labels will form a compact cluster in this graph. We recover the clusters using a spectral theory based technique. The results are demonstrated on the Corel5k dataset. With improvements in the range of 25%-55% the performance in some cases approaches the state of the art despite using a very simple base algorithm. |
| doi_str_mv | 10.1109/ICIP.2009.5414229 |
| format | Conference Proceeding |
| fullrecord | <record><control><sourceid>ieee_6IE</sourceid><recordid>TN_cdi_ieee_primary_5414229</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>5414229</ieee_id><sourcerecordid>5414229</sourcerecordid><originalsourceid>FETCH-LOGICAL-i175t-2bc82a1db989a34bd9e64434401634997bdd97519bec8e0729ca7d866227d2923</originalsourceid><addsrcrecordid>eNpVkM9KAzEYxOM_cK19APGSF9iafEk2-Y5lsXWhoAc9l2STrZHubtlE0Le3aC-eZpgfDMMQcsfZgnOGD03dvCyAMVwoySUAnpE5anO0UqpKKXZOChCGl0ZJvPjHRHVJCq4ASmkMuyY3KX0wBowLXpD1kqZDaPNk97QP-X30tBsn2o5DDl-ZOpuCpz4m27u4-7Q5jgMdOxp7uwvUDsOYf7N0S646u09hftIZeVs9vtZP5eZ53dTLTRm5VrkE1xqw3Ds0aIV0HkMlpZCS8UpIRO28R604utCawDRga7U3VQWgPSCIGbn_640hhO1hOg6ZvrenT8QPihNQPQ</addsrcrecordid><sourcetype>Publisher</sourcetype><iscdi>true</iscdi><recordtype>conference_proceeding</recordtype></control><display><type>conference_proceeding</type><title>A spectral method for context based disambiguation of image annotations</title><source>IEEE Electronic Library (IEL) Conference Proceedings</source><creator>Semenovich, D. ; Sowmya, A.</creator><creatorcontrib>Semenovich, D. ; Sowmya, A.</creatorcontrib><description>In this work we employ contextual information to improve the quality of image labellings provided by an existing automatic image annotation algorithm in a weakly supervised setting, where each training image is labelled but it is not known which part of the image its labels are referring to. We recast the problem into that of constructing a graph which encodes pairwise consistency of candidate annotations and observe that mutually consistent labels will form a compact cluster in this graph. We recover the clusters using a spectral theory based technique. The results are demonstrated on the Corel5k dataset. With improvements in the range of 25%-55% the performance in some cases approaches the state of the art despite using a very simple base algorithm.</description><identifier>ISSN: 1522-4880</identifier><identifier>ISBN: 9781424456536</identifier><identifier>ISBN: 1424456533</identifier><identifier>EISSN: 2381-8549</identifier><identifier>EISBN: 9781424456550</identifier><identifier>EISBN: 9781424456543</identifier><identifier>EISBN: 142445655X</identifier><identifier>EISBN: 1424456541</identifier><identifier>DOI: 10.1109/ICIP.2009.5414229</identifier><language>eng</language><publisher>IEEE</publisher><subject>Australia ; Clustering algorithms ; Computer science ; Context modeling ; Image annotation ; Image recognition ; Image retrieval ; Information retrieval ; Labeling ; Layout ; Machine learning ; spectral clustering</subject><ispartof>2009 16th IEEE International Conference on Image Processing (ICIP), 2009, p.789-792</ispartof><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/5414229$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>309,310,780,784,789,790,2058,27925,54920</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/5414229$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Semenovich, D.</creatorcontrib><creatorcontrib>Sowmya, A.</creatorcontrib><title>A spectral method for context based disambiguation of image annotations</title><title>2009 16th IEEE International Conference on Image Processing (ICIP)</title><addtitle>ICIP</addtitle><description>In this work we employ contextual information to improve the quality of image labellings provided by an existing automatic image annotation algorithm in a weakly supervised setting, where each training image is labelled but it is not known which part of the image its labels are referring to. We recast the problem into that of constructing a graph which encodes pairwise consistency of candidate annotations and observe that mutually consistent labels will form a compact cluster in this graph. We recover the clusters using a spectral theory based technique. The results are demonstrated on the Corel5k dataset. With improvements in the range of 25%-55% the performance in some cases approaches the state of the art despite using a very simple base algorithm.</description><subject>Australia</subject><subject>Clustering algorithms</subject><subject>Computer science</subject><subject>Context modeling</subject><subject>Image annotation</subject><subject>Image recognition</subject><subject>Image retrieval</subject><subject>Information retrieval</subject><subject>Labeling</subject><subject>Layout</subject><subject>Machine learning</subject><subject>spectral clustering</subject><issn>1522-4880</issn><issn>2381-8549</issn><isbn>9781424456536</isbn><isbn>1424456533</isbn><isbn>9781424456550</isbn><isbn>9781424456543</isbn><isbn>142445655X</isbn><isbn>1424456541</isbn><fulltext>true</fulltext><rsrctype>conference_proceeding</rsrctype><creationdate>2009</creationdate><recordtype>conference_proceeding</recordtype><sourceid>6IE</sourceid><sourceid>RIE</sourceid><recordid>eNpVkM9KAzEYxOM_cK19APGSF9iafEk2-Y5lsXWhoAc9l2STrZHubtlE0Le3aC-eZpgfDMMQcsfZgnOGD03dvCyAMVwoySUAnpE5anO0UqpKKXZOChCGl0ZJvPjHRHVJCq4ASmkMuyY3KX0wBowLXpD1kqZDaPNk97QP-X30tBsn2o5DDl-ZOpuCpz4m27u4-7Q5jgMdOxp7uwvUDsOYf7N0S646u09hftIZeVs9vtZP5eZ53dTLTRm5VrkE1xqw3Ds0aIV0HkMlpZCS8UpIRO28R604utCawDRga7U3VQWgPSCIGbn_640hhO1hOg6ZvrenT8QPihNQPQ</recordid><startdate>200911</startdate><enddate>200911</enddate><creator>Semenovich, D.</creator><creator>Sowmya, A.</creator><general>IEEE</general><scope>6IE</scope><scope>6IH</scope><scope>CBEJK</scope><scope>RIE</scope><scope>RIO</scope></search><sort><creationdate>200911</creationdate><title>A spectral method for context based disambiguation of image annotations</title><author>Semenovich, D. ; Sowmya, A.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-i175t-2bc82a1db989a34bd9e64434401634997bdd97519bec8e0729ca7d866227d2923</frbrgroupid><rsrctype>conference_proceedings</rsrctype><prefilter>conference_proceedings</prefilter><language>eng</language><creationdate>2009</creationdate><topic>Australia</topic><topic>Clustering algorithms</topic><topic>Computer science</topic><topic>Context modeling</topic><topic>Image annotation</topic><topic>Image recognition</topic><topic>Image retrieval</topic><topic>Information retrieval</topic><topic>Labeling</topic><topic>Layout</topic><topic>Machine learning</topic><topic>spectral clustering</topic><toplevel>online_resources</toplevel><creatorcontrib>Semenovich, D.</creatorcontrib><creatorcontrib>Sowmya, A.</creatorcontrib><collection>IEEE Electronic Library (IEL) Conference Proceedings</collection><collection>IEEE Proceedings Order Plan (POP) 1998-present by volume</collection><collection>IEEE Xplore All Conference Proceedings</collection><collection>IEEE Xplore</collection><collection>IEEE Proceedings Order Plans (POP) 1998-present</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Semenovich, D.</au><au>Sowmya, A.</au><format>book</format><genre>proceeding</genre><ristype>CONF</ristype><atitle>A spectral method for context based disambiguation of image annotations</atitle><btitle>2009 16th IEEE International Conference on Image Processing (ICIP)</btitle><stitle>ICIP</stitle><date>2009-11</date><risdate>2009</risdate><spage>789</spage><epage>792</epage><pages>789-792</pages><issn>1522-4880</issn><eissn>2381-8549</eissn><isbn>9781424456536</isbn><isbn>1424456533</isbn><eisbn>9781424456550</eisbn><eisbn>9781424456543</eisbn><eisbn>142445655X</eisbn><eisbn>1424456541</eisbn><abstract>In this work we employ contextual information to improve the quality of image labellings provided by an existing automatic image annotation algorithm in a weakly supervised setting, where each training image is labelled but it is not known which part of the image its labels are referring to. We recast the problem into that of constructing a graph which encodes pairwise consistency of candidate annotations and observe that mutually consistent labels will form a compact cluster in this graph. We recover the clusters using a spectral theory based technique. The results are demonstrated on the Corel5k dataset. With improvements in the range of 25%-55% the performance in some cases approaches the state of the art despite using a very simple base algorithm.</abstract><pub>IEEE</pub><doi>10.1109/ICIP.2009.5414229</doi><tpages>4</tpages></addata></record> |
| fulltext | fulltext_linktorsrc |
| identifier | ISSN: 1522-4880 |
| ispartof | 2009 16th IEEE International Conference on Image Processing (ICIP), 2009, p.789-792 |
| issn | 1522-4880 2381-8549 |
| language | eng |
| recordid | cdi_ieee_primary_5414229 |
| source | IEEE Electronic Library (IEL) Conference Proceedings |
| subjects | Australia Clustering algorithms Computer science Context modeling Image annotation Image recognition Image retrieval Information retrieval Labeling Layout Machine learning spectral clustering |
| title | A spectral method for context based disambiguation of image annotations |
| url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-01T06%3A54%3A59IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-ieee_6IE&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=proceeding&rft.atitle=A%20spectral%20method%20for%20context%20based%20disambiguation%20of%20image%20annotations&rft.btitle=2009%2016th%20IEEE%20International%20Conference%20on%20Image%20Processing%20(ICIP)&rft.au=Semenovich,%20D.&rft.date=2009-11&rft.spage=789&rft.epage=792&rft.pages=789-792&rft.issn=1522-4880&rft.eissn=2381-8549&rft.isbn=9781424456536&rft.isbn_list=1424456533&rft_id=info:doi/10.1109/ICIP.2009.5414229&rft_dat=%3Cieee_6IE%3E5414229%3C/ieee_6IE%3E%3Curl%3E%3C/url%3E&rft.eisbn=9781424456550&rft.eisbn_list=9781424456543&rft.eisbn_list=142445655X&rft.eisbn_list=1424456541&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=5414229&rfr_iscdi=true |