Learning from Weak and Noisy Labels for Semantic Segmentation

A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels. By avoiding the tedious pixel-level annotation process, it can exploit the unlimited supply of user-tagged images from media-sharing sites suc...

Full description

Saved in:

Bibliographic Details
Published in:	IEEE transactions on pattern analysis and machine intelligence Vol. 39; no. 3; pp. 486 - 500
Main Authors:	Lu, Zhiwu, Fu, Zhenyong, Xiang, Tao, Han, Peng, Wang, Liwei, Gao, Xin
Format:	Journal Article
Language:	English
Published:	United States IEEE 01.03.2017 The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Subjects:	Algorithms Computational modeling Image annotation Image segmentation label noise reduction Labeling Labels Machine learning Noise measurement Noise reduction Optimization Pixels Semantic segmentation Semantics sparse learning Training weakly supervised learning
ISSN:	0162-8828, 2160-9292, 1939-3539
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Abstract	A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels. By avoiding the tedious pixel-level annotation process, it can exploit the unlimited supply of user-tagged images from media-sharing sites such as Flickr for large scale applications. However, these `free' tags/labels are often noisy and few existing works address the problem of learning with both weak and noisy labels. In this work, we cast the WSSS problem into a label noise reduction problem. Specifically, after segmenting each image into a set of superpixels, the weak and potentially noisy image-level labels are propagated to the superpixel level resulting in highly noisy labels; the key to semantic segmentation is thus to identify and correct the superpixel noisy labels. To this end, a novel L 1 -optimisation based sparse learning model is formulated to directly and explicitly detect noisy labels. To solve the L 1 -optimisation problem, we further develop an efficient learning algorithm by introducing an intermediate labelling variable. Extensive experiments on three benchmark datasets show that our method yields state-of-the-art results given noise-free labels, whilst significantly outperforming the existing methods when the weak labels are also noisy.
AbstractList	A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels. By avoiding the tedious pixel-level annotation process, it can exploit the unlimited supply of user-tagged images from media-sharing sites such as Flickr for large scale applications. However, these `free' tags/labels are often noisy and few existing works address the problem of learning with both weak and noisy labels. In this work, we cast the WSSS problem into a label noise reduction problem. Specifically, after segmenting each image into a set of superpixels, the weak and potentially noisy image-level labels are propagated to the superpixel level resulting in highly noisy labels; the key to semantic segmentation is thus to identify and correct the superpixel noisy labels. To this end, a novel L -optimisation based sparse learning model is formulated to directly and explicitly detect noisy labels. To solve the L -optimisation problem, we further develop an efficient learning algorithm by introducing an intermediate labelling variable. Extensive experiments on three benchmark datasets show that our method yields state-of-the-art results given noise-free labels, whilst significantly outperforming the existing methods when the weak labels are also noisy. A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels. By avoiding the tedious pixel-level annotation process, it can exploit the unlimited supply of user-tagged images from media-sharing sites such as Flickr for large scale applications. However, these `free' tags/labels are often noisy and few existing works address the problem of learning with both weak and noisy labels. In this work, we cast the WSSS problem into a label noise reduction problem. Specifically, after segmenting each image into a set of superpixels, the weak and potentially noisy image-level labels are propagated to the superpixel level resulting in highly noisy labels; the key to semantic segmentation is thus to identify and correct the superpixel noisy labels. To this end, a novel L1-optimisation based sparse learning model is formulated to directly and explicitly detect noisy labels. To solve the L1-optimisation problem, we further develop an efficient learning algorithm by introducing an intermediate labelling variable. Extensive experiments on three benchmark datasets show that our method yields state-of-the-art results given noise-free labels, whilst significantly outperforming the existing methods when the weak labels are also noisy. A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels. By avoiding the tedious pixel-level annotation process, it can exploit the unlimited supply of user-tagged images from media-sharing sites such as Flickr for large scale applications. However, these `free' tags/labels are often noisy and few existing works address the problem of learning with both weak and noisy labels. In this work, we cast the WSSS problem into a label noise reduction problem. Specifically, after segmenting each image into a set of superpixels, the weak and potentially noisy image-level labels are propagated to the superpixel level resulting in highly noisy labels; the key to semantic segmentation is thus to identify and correct the superpixel noisy labels. To this end, a novel L 1 -optimisation based sparse learning model is formulated to directly and explicitly detect noisy labels. To solve the L 1 -optimisation problem, we further develop an efficient learning algorithm by introducing an intermediate labelling variable. Extensive experiments on three benchmark datasets show that our method yields state-of-the-art results given noise-free labels, whilst significantly outperforming the existing methods when the weak labels are also noisy.
Author	Peng Han Zhiwu Lu Liwei Wang Xin Gao Tao Xiang Zhenyong Fu
Author_xml	– sequence: 1 givenname: Zhiwu surname: Lu fullname: Lu, Zhiwu – sequence: 2 givenname: Zhenyong surname: Fu fullname: Fu, Zhenyong – sequence: 3 givenname: Tao orcidid: 0000-0002-2530-1059 surname: Xiang fullname: Xiang, Tao – sequence: 4 givenname: Peng surname: Han fullname: Han, Peng – sequence: 5 givenname: Liwei surname: Wang fullname: Wang, Liwei – sequence: 6 givenname: Xin surname: Gao fullname: Gao, Xin
BackLink	https://www.ncbi.nlm.nih.gov/pubmed/28113885$$D View this record in MEDLINE/PubMed
BookMark	eNp9kU1P3DAQhi1EBQvlD4BUReLSSxaPHX_k0MMK0YK0LZW6VY-R44yRIbHBzh749812Fw576Mkj-XlnRs-ckMMQAxJyDnQOQOur1c_F97s5oyDnTAgGih2QGQNJy5rV7JDMph9Was30MTnJ-ZFSqATlR-SYaQCutZiRL0s0KfjwULgUh-IPmqfChK74EX1-LZamxT4XLqbiFw4mjN5OxcOAYTSjj-Ej-eBMn_Fs956S319vVte35fL-2931YlnaCtRYOimZrLoOOXBUympuAa1QrXHC1ZJSzirbdugUyE4jKitqgU5QACucbvkp-bzt-5ziyxrz2Aw-W-x7EzCucwNaggTOmZrQyz30Ma5TmLbbUJKruq7kRH3aUet2wK55Tn4w6bV5EzMBbAvYFHNO6N4RoM3GfvPPfrOx3-zsTyG9F7J-K2pMxvf_j15sox4R32ep6V6gFP8LqPGQHg
CODEN	ITPIDJ
CitedBy_id	crossref_primary_10_1016_j_media_2021_102214 crossref_primary_10_1109_ACCESS_2019_2905879 crossref_primary_10_1109_TGRS_2018_2823866 crossref_primary_10_1109_TGRS_2018_2848725 crossref_primary_10_1109_TIP_2018_2881928 crossref_primary_10_1109_TMM_2019_2955240 crossref_primary_10_1016_j_neunet_2021_06_012 crossref_primary_10_1016_j_cageo_2021_104843 crossref_primary_10_1016_j_asoc_2024_112138 crossref_primary_10_1145_3480970 crossref_primary_10_1145_3572916 crossref_primary_10_1007_s10994_021_06008_4 crossref_primary_10_1007_s10278_025_01408_7 crossref_primary_10_1109_TPAMI_2020_3047817 crossref_primary_10_1007_s10462_019_09792_7 crossref_primary_10_1109_TGRS_2021_3068280 crossref_primary_10_1109_TNNLS_2018_2792062 crossref_primary_10_1016_j_cmpb_2019_05_010 crossref_primary_10_1109_TIP_2021_3089943 crossref_primary_10_1007_s10707_021_00459_6 crossref_primary_10_3390_rs14071561 crossref_primary_10_1109_ACCESS_2020_2989200 crossref_primary_10_3390_rs12203421 crossref_primary_10_1016_j_bspc_2023_105473 crossref_primary_10_1109_TPAMI_2018_2852750 crossref_primary_10_1016_j_neuroimage_2021_118568 crossref_primary_10_1080_01431161_2021_1973685 crossref_primary_10_1109_TNNLS_2021_3067107 crossref_primary_10_1109_TIP_2019_2936649 crossref_primary_10_1007_s11554_018_0762_3 crossref_primary_10_3390_app122111226 crossref_primary_10_1007_s11042_023_15983_w crossref_primary_10_1016_j_knosys_2021_106771 crossref_primary_10_3389_fnins_2021_610122 crossref_primary_10_1016_j_neucom_2018_10_061 crossref_primary_10_1016_j_bdr_2021_100272 crossref_primary_10_1109_JSTARS_2020_2994162 crossref_primary_10_3390_rs11232823 crossref_primary_10_1016_j_eswa_2022_117030 crossref_primary_10_1016_j_swevo_2017_11_003 crossref_primary_10_1109_TIP_2018_2836306 crossref_primary_10_1016_j_isprsjprs_2023_01_021 crossref_primary_10_1007_s11042_020_09730_8 crossref_primary_10_1007_s11263_021_01553_w crossref_primary_10_1007_s00521_021_06378_9 crossref_primary_10_1109_LGRS_2018_2842792 crossref_primary_10_1109_TIP_2018_2877939 crossref_primary_10_1109_TPAMI_2019_2941684 crossref_primary_10_1109_TGRS_2019_2961141 crossref_primary_10_1190_geo2018_0028_1 crossref_primary_10_1016_j_engappai_2020_103708 crossref_primary_10_1016_j_asoc_2018_10_035 crossref_primary_10_3390_rs16122080 crossref_primary_10_1016_j_patcog_2021_108467 crossref_primary_10_3390_math9192498 crossref_primary_10_1109_TGRS_2023_3264232 crossref_primary_10_3390_app13137966 crossref_primary_10_1109_TGRS_2018_2867444 crossref_primary_10_1155_2019_9180391 crossref_primary_10_1109_TMM_2020_2991592
Cites_doi	10.1109/TIP.2007.911828 10.1109/CVPR.2010.5540060 10.1109/CVPR.2012.6247719 10.1109/CVPR.2014.415 10.1137/090777761 10.1007/s11263-012-0574-z 10.1109/CVPR.2007.383098 10.1109/TNNLS.2013.2292894 10.1109/CVPR.2015.7298888 10.1109/CVPR.2015.7299002 10.1007/978-3-319-10599-4_33 10.1109/CVPR.2015.7298965 10.1109/CVPR.2015.7298780 10.5244/C.29.29 10.1109/CVPR.2008.4587503 10.1109/TPAMI.2012.231 10.1145/1631272.1631305 10.1109/ICIP.2012.6467501 10.1007/s11263-010-0344-8 10.1109/CVPR.2012.6247757 10.1109/CVPR.2013.386 10.1109/TPAMI.2008.79 10.1145/3065386 10.1007/s11263-009-0245-x 10.1109/ICCV.2011.6126299 10.1109/TMM.2013.2285526 10.1109/CVPR.2014.53 10.1109/CVPR.2014.81 10.1109/CVPR.2014.408 10.1007/978-3-319-10584-0_28 10.1109/TPAMI.2011.131 10.1109/CVPR.2014.190 10.1007/s11263-011-0449-8 10.1109/CVPR.2013.270 10.1109/CVPR.2014.310 10.1109/TMM.2011.2174780 10.1109/CVPR.2014.119 10.1109/TIP.2006.881969 10.1145/1631272.1631291 10.1109/ICCV.2011.6126219 10.1007/s11263-008-0202-0 10.1109/TPAMI.2015.2456887 10.1109/CVPR.2014.49 10.1109/ICCV.2009.5459248
ContentType	Journal Article
Copyright	Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2017
Copyright_xml	– notice: Copyright The Institute of Electrical and Electronics Engineers, Inc. (IEEE) 2017
DBID	97E RIA RIE AAYXX CITATION NPM 7SC 7SP 8FD JQ2 L7M L~C L~D 7X8
DOI	10.1109/TPAMI.2016.2552172
DatabaseName	IEEE All-Society Periodicals Package (ASPP) 2005–Present IEEE All-Society Periodicals Package (ASPP) 1998–Present IEEE Electronic Library (IEL) CrossRef PubMed Computer and Information Systems Abstracts Electronics & Communications Abstracts Technology Research Database ProQuest Computer Science Collection Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Academic Computer and Information Systems Abstracts Professional MEDLINE - Academic
DatabaseTitle	CrossRef PubMed Technology Research Database Computer and Information Systems Abstracts – Academic Electronics & Communications Abstracts ProQuest Computer Science Collection Computer and Information Systems Abstracts Advanced Technologies Database with Aerospace Computer and Information Systems Abstracts Professional MEDLINE - Academic
DatabaseTitleList	PubMed Technology Research Database MEDLINE - Academic
Database_xml	– sequence: 1 dbid: NPM name: PubMed url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed sourceTypes: Index Database – sequence: 2 dbid: RIE name: IEEE Electronic Library (IEL) url: https://ieeexplore.ieee.org/ sourceTypes: Publisher – sequence: 3 dbid: 7X8 name: MEDLINE - Academic url: https://search.proquest.com/medline sourceTypes: Aggregation Database
DeliveryMethod	fulltext_linktorsrc
Discipline	Engineering Computer Science
EISSN	2160-9292 1939-3539
EndPage	500
ExternalDocumentID	28113885 10_1109_TPAMI_2016_2552172 7450177
Genre	orig-research Research Support, Non-U.S. Gov't Journal Article
GrantInformation_xml	– fundername: 973 Program of China grantid: 2014CB340403; 2015CB352502 – fundername: KAUST funderid: 10.13039/501100004052 – fundername: Fundamental Research Funds for the Central Universities and the Research Funds of Renmin University of China grantid: 15XNLQ01 – fundername: European Research Council FP7 Project SUNNY grantid: 313243 funderid: 10.13039/501100000781 – fundername: National Natural Science Foundation of China grantid: 61573363; 61573026 funderid: 10.13039/501100001809 – fundername: IBM Global SUR Award Program funderid: 10.13039/100004316
GroupedDBID	--- -DZ -~X .DC 0R~ 29I 4.4 53G 5GY 6IK 97E AAJGR AARMG AASAJ AAWTH ABAZT ABQJQ ABVLG ACGFO ACGFS ACIWK ACNCT AENEX AGQYO AHBIQ AKJIK AKQYR ALMA_UNASSIGNED_HOLDINGS ASUFR ATWAV BEFXN BFFAM BGNUA BKEBE BPEOZ CS3 DU5 E.L EBS EJD F5P HZ~ IEDLZ IFIPE IPLJI JAVBF LAI M43 MS~ O9- OCL P2P PQQKQ RIA RIE RNS RXW TAE TN5 UHB ~02 AAYXX CITATION 5VS 9M8 ABFSI ADRHT AETEA AETIX AGSQL AI. AIBXA ALLEH FA8 H~9 IBMZZ ICLAB IFJZH NPM RIG RNI RZB VH1 XJT 7SC 7SP 8FD JQ2 L7M L~C L~D 7X8
ID	FETCH-LOGICAL-c417t-f66264dde313e77c83c1ec57baf5f9600324cbdef716d8ee7c595ef5011c5f8b3
IEDL.DBID	RIE
ISICitedReferencesCount	89
ISICitedReferencesURI	http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000395555100006&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN	0162-8828
IngestDate	Sun Sep 28 16:18:21 EDT 2025 Sun Nov 30 04:07:21 EST 2025 Mon Jul 21 05:42:28 EDT 2025 Sat Nov 29 05:15:57 EST 2025 Tue Nov 18 22:18:34 EST 2025 Wed Aug 27 02:47:49 EDT 2025
IsDoiOpenAccess	false
IsOpenAccess	true
IsPeerReviewed	true
IsScholarly	true
Issue	3
Language	English
License	https://ieeexplore.ieee.org/Xplorehelp/downloads/license-information/IEEE.html
LinkModel	DirectLink
MergedId	FETCHMERGED-LOGICAL-c417t-f66264dde313e77c83c1ec57baf5f9600324cbdef716d8ee7c595ef5011c5f8b3
Notes	ObjectType-Article-1 SourceType-Scholarly Journals-1 ObjectType-Feature-2 content type line 14 content type line 23
ORCID	0000-0002-2530-1059
PMID	28113885
PQID	1866379946
PQPubID	85458
PageCount	15
ParticipantIDs	ieee_primary_7450177 crossref_primary_10_1109_TPAMI_2016_2552172 proquest_miscellaneous_1861613327 pubmed_primary_28113885 crossref_citationtrail_10_1109_TPAMI_2016_2552172 proquest_journals_1866379946
PublicationCentury	2000
PublicationDate	2017-03-01
PublicationDateYYYYMMDD	2017-03-01
PublicationDate_xml	– month: 03 year: 2017 text: 2017-03-01 day: 01
PublicationDecade	2010
PublicationPlace	United States
PublicationPlace_xml	– name: United States – name: New York
PublicationTitle	IEEE transactions on pattern analysis and machine intelligence
PublicationTitleAbbrev	TPAMI
PublicationTitleAlternate	IEEE Trans Pattern Anal Mach Intell
PublicationYear	2017
Publisher	IEEE The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Publisher_xml	– name: IEEE – name: The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
References	ref13 ref12 hariharan (ref37) 0 ref58 ref14 ref53 ref52 ref11 fergus (ref50) 0 ref10 chen (ref38) 2015 ref17 ref16 ref19 lucchi (ref63) 0 ref18 rubinstein (ref66) 0 ladicky (ref5) 0 ref51 gould (ref8) 0 sermanet (ref55) 0 ref48 ref47 ref42 ref44 li (ref33) 2009 ref43 everingham (ref56) 2007 xie (ref59) 0 shotton (ref57) 0 ref7 ref9 ref4 ref3 ref6 ref40 ref35 ref34 ref36 ref31 ref30 ref32 ref2 zhu (ref1) 2015 lee (ref25) 0 zhu (ref45) 0 zhou (ref46) 0 ref68 ref24 ref67 ref23 ref20 chen (ref26) 0 ref22 ref65 ref21 liu (ref49) 0 zhang (ref15) 0 ref28 ref27 ref29 pathak (ref39) 0 deng (ref54) 0 yao (ref64) 0 ref60 ref62 ref61 yuan (ref41) 0
References_xml	– ident: ref23 doi: 10.1109/TIP.2007.911828 – ident: ref10 doi: 10.1109/CVPR.2010.5540060 – ident: ref40 doi: 10.1109/CVPR.2012.6247719 – ident: ref6 doi: 10.1109/CVPR.2014.415 – ident: ref48 doi: 10.1137/090777761 – year: 2015 ident: ref38 article-title: Semantic image segmentation with deep convolutional nets and fully connected CRFs publication-title: ICLRE – ident: ref67 doi: 10.1007/s11263-012-0574-z – start-page: 105 year: 0 ident: ref26 article-title: Smoothing proximal gradient method for general structured sparse learning publication-title: Proc Conf Annu Conf Uncertainty Artif Intell – ident: ref9 doi: 10.1109/CVPR.2007.383098 – year: 0 ident: ref55 article-title: OverFeat: Integrated recognition, localization and detection using convolutional networks publication-title: ICLRE – ident: ref44 doi: 10.1109/TNNLS.2013.2292894 – ident: ref19 doi: 10.1109/CVPR.2015.7298888 – start-page: 2853 year: 0 ident: ref59 article-title: Semantic graph construction for weakly-supervised image parsing publication-title: Proc 28th Nat Conf Artif Intell – ident: ref18 doi: 10.1109/CVPR.2015.7299002 – ident: ref7 doi: 10.1007/978-3-319-10599-4_33 – start-page: 912 year: 0 ident: ref45 article-title: Semi-supervised learning using Gaussian fields and harmonic functions publication-title: Proc Int Conf Mach Learn – start-page: 663 year: 0 ident: ref49 article-title: Robust subspace segmentation by low-rank representation publication-title: Proc Int Conf Mach Learn – start-page: 297 year: 0 ident: ref37 article-title: Simultaneous detection and segmentation publication-title: Proc 13th Eur Conf Comput Vis – ident: ref31 doi: 10.1109/CVPR.2015.7298965 – ident: ref32 doi: 10.1109/CVPR.2015.7298780 – ident: ref65 doi: 10.5244/C.29.29 – ident: ref2 doi: 10.1109/CVPR.2008.4587503 – start-page: 1 year: 0 ident: ref57 article-title: Textonboost: Joint appearance, shape and context modeling for multi-class object recognition and segmentation publication-title: Proc 9th Eur Conf Comput Vis – ident: ref36 doi: 10.1109/TPAMI.2012.231 – year: 0 ident: ref39 article-title: Fully convolutional multi-class multiple instance learning publication-title: ICLRE – ident: ref20 doi: 10.1145/1631272.1631305 – start-page: 695 year: 0 ident: ref41 article-title: A novel topic-level random walk framework for scene image co-segmentation publication-title: Proc 13th Eur Conf Comput Vis – start-page: 702 year: 0 ident: ref64 article-title: Describing the scene as a whole: Joint object detection, scene classification and semantic segmentation publication-title: Proc IEEE Conf Comput Vis Pattern Recog – ident: ref43 doi: 10.1109/ICIP.2012.6467501 – ident: ref27 doi: 10.1007/s11263-010-0344-8 – start-page: 400 year: 0 ident: ref63 article-title: Structured image segmentation using kernelized features publication-title: Proc 12th Eur Conf Comput Vis – ident: ref12 doi: 10.1109/CVPR.2012.6247757 – ident: ref68 doi: 10.1109/CVPR.2013.386 – ident: ref24 doi: 10.1109/TPAMI.2008.79 – ident: ref35 doi: 10.1145/3065386 – start-page: 1889 year: 0 ident: ref15 article-title: Sparse reconstruction for weakly supervised semantic segmentation publication-title: Proc 23rd Int Joint Conf Artif Intell – ident: ref61 doi: 10.1007/s11263-009-0245-x – start-page: 2036 year: 2009 ident: ref33 article-title: Towards total scene understanding: Classification, annotation and segmentation in an automatic framework. publication-title: Proc IEEE Conf Comput Vis Pattern Recog – ident: ref11 doi: 10.1109/ICCV.2011.6126299 – ident: ref16 doi: 10.1109/TMM.2013.2285526 – start-page: 632 year: 0 ident: ref8 article-title: Superpixel graph label transfer with learned distance metric publication-title: Proc 13th Eur Conf Comput Vis – ident: ref30 doi: 10.1109/CVPR.2014.53 – ident: ref53 doi: 10.1109/CVPR.2014.81 – ident: ref17 doi: 10.1109/CVPR.2014.408 – ident: ref21 doi: 10.1007/978-3-319-10584-0_28 – year: 2007 ident: ref56 article-title: The PASCAL visual object classes challenge 2007 (VOC2007) Results – ident: ref58 doi: 10.1109/TPAMI.2011.131 – start-page: 321 year: 0 ident: ref46 article-title: Learning with local and global consistency publication-title: Proc Adv Neural Inf Process Syst – ident: ref42 doi: 10.1109/CVPR.2014.190 – ident: ref62 doi: 10.1007/s11263-011-0449-8 – ident: ref29 doi: 10.1007/s11263-012-0574-z – ident: ref14 doi: 10.1109/CVPR.2013.270 – ident: ref52 doi: 10.1109/CVPR.2014.310 – ident: ref13 doi: 10.1109/TMM.2011.2174780 – ident: ref34 doi: 10.1109/CVPR.2014.119 – ident: ref22 doi: 10.1109/TIP.2006.881969 – ident: ref60 doi: 10.1145/1631272.1631291 – start-page: 85 year: 0 ident: ref66 article-title: Annotation propagation in large image databases via dense image correspondence publication-title: Proc 12th Eur Conf Comput Vis – ident: ref28 doi: 10.1109/ICCV.2011.6126219 – start-page: 248 year: 0 ident: ref54 article-title: ImageNet: A large-scale hierarchical image database publication-title: Proc IEEE Conf Comput Vis Pattern Recog – ident: ref3 doi: 10.1007/s11263-008-0202-0 – year: 2015 ident: ref1 article-title: Beyond pixels: A comprehensive survey from bottom-up to semantic image segmentation and cosegmentation publication-title: CoRR – ident: ref47 doi: 10.1109/TPAMI.2015.2456887 – ident: ref51 doi: 10.1109/CVPR.2014.49 – start-page: 522 year: 0 ident: ref50 article-title: Semi-supervised learning in gigantic image collections publication-title: Proc Adv Neural Inf Process Syst – start-page: 801 year: 0 ident: ref25 article-title: Efficient sparse coding algorithms publication-title: Proc Adv Neural Inf Process Syst – ident: ref4 doi: 10.1109/ICCV.2009.5459248 – start-page: 239 year: 0 ident: ref5 article-title: Graph cut based inference with co-occurrence statistics publication-title: Proc 11th Eur Conf Comput Vis
SSID	ssj0014503
Score	2.5544004
Snippet	A weakly supervised semantic segmentation (WSSS) method aims to learn a segmentation model from weak (image-level) as opposed to strong (pixel-level) labels....
SourceID	proquest pubmed crossref ieee
SourceType	Aggregation Database Index Database Enrichment Source Publisher
StartPage	486
SubjectTerms	Algorithms Computational modeling Image annotation Image segmentation label noise reduction Labeling Labels Machine learning Noise measurement Noise reduction Optimization Pixels Semantic segmentation Semantics sparse learning Training weakly supervised learning
Title	Learning from Weak and Noisy Labels for Semantic Segmentation
URI	https://ieeexplore.ieee.org/document/7450177 https://www.ncbi.nlm.nih.gov/pubmed/28113885 https://www.proquest.com/docview/1866379946 https://www.proquest.com/docview/1861613327
Volume	39
WOSCitedRecordID	wos000395555100006&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText	1
inHoldings	1
isFullTextHit
isPrint
journalDatabaseRights	– providerCode: PRVIEE databaseName: IEEE Electronic Library (IEL) customDbUrl: eissn: 2160-9292 dateEnd: 99991231 omitProxy: false ssIdentifier: ssj0014503 issn: 0162-8828 databaseCode: RIE dateStart: 19790101 isFulltext: true titleUrlDefault: https://ieeexplore.ieee.org/ providerName: IEEE
link	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV1NT9wwEB3Bqgc4dFs-2m0pcqXeaCBeO3F86AFVRa3UrpAKYm-RY09WK0oWsbtI_fcdO07EoSBxi2THTjwzmZnY8x7AJ8clcmN4gm6sE-nyKqHMVieVQnIQng5JVoFsQk0mxXSqzzfgc18Lg4jh8Bke-8uwl-8Wdu1_lZ0omZECqU3YVCpva7X6HQNqEi2ON1k4TdIVyKT65OL89NcPf4orP6YA2jMyeQjggnNReArlB_4oEKw8HmsGn3M2fN7TvoKXMbZkp60yvIYNbHZg2PE2sGjGO7D9AIRwF75EiNUZ86Um7ArNNTONY5PFfPmX_TQVeU9GoS37jTckhrmli9lNLFlq9uDy7NvF1-9JJFVIrORqldQ5pTCSPmqCC1TKFsJytJmqTJ3VlM6kFGHZymFNiZQrEJXNdIY1vQq3WV1UYh8GzaLBt8BqYZxVunCUs8ixscZRbpRalLVNHXfVCHi3tKWNiOOe-OJPGTKPVJdBMqWXTBklM4Kj_p7bFm_jyd67ft37nnHJR3DQSbCMJrksPbKfUFrLfAQf-2YyJr9DYhpcrEMfioCFGNMQb1rJ92N3CvPu_3O-h62x9_jheNoBDFZ3a_wAL-z9ar68OySNnRaHQWP_AQcK47k
linkProvider	IEEE
linkToHtml	http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwlV3BTtwwEB1RWqn0UFqgdFtoXam3NhDHzjo-9IAqEKjLCqlblVvk2BO0KmQRu4vE33fsOBGHFolbJDt24pnJzIs98wA-Oy6RG8MTdJlOpBtWCSFbnVQKyUF4OiRZBbIJNR4X5-f6bAW-9rkwiBgOn-Gevwx7-W5ml_5X2b6SOSmQegJPcymztM3W6vcMqFG0lbzJxmmaLkUm1fuTs4PTE3-Oa7hHIbTnZPJFgAvOReFJlO95pECx8v9oM3ido_XHPe8reBmjS3bQqsNrWMFmA9Y75gYWDXkDXtwrQ7gJ32KR1Qvmk03YbzR_mGkcG8-m8zs2MhX5T0bBLfuJVySIqaWLi6uYtNRswa-jw8n34yTSKiRWcrVI6iGBGEmfNcEFKmULYTnaXFWmzmsCNCnFWLZyWBOUcgWisrnOsaZX4Tavi0q8gdVm1uBbYLUwzipdOEItMjPWOEJHqUVZ29RxVw2Ad0tb2lhz3FNfXJYBe6S6DJIpvWTKKJkBfOnvuW4rbjzYe9Ove98zLvkAdjoJltEo56Wv7SeU1nI4gE99M5mT3yMxDc6WoQ_FwEJkNMR2K_l-7E5h3v17zo_w_HhyOipHJ-Mf72Et8_4_HFbbgdXFzRJ34Zm9XUznNx-C3v4FqOnmGA
openUrl	ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Learning+from+Weak+and+Noisy+Labels+for+Semantic+Segmentation&rft.jtitle=IEEE+transactions+on+pattern+analysis+and+machine+intelligence&rft.au=Lu%2C+Zhiwu&rft.au=Fu%2C+Zhenyong&rft.au=Xiang%2C+Tao&rft.au=Han%2C+Peng&rft.date=2017-03-01&rft.pub=The+Institute+of+Electrical+and+Electronics+Engineers%2C+Inc.+%28IEEE%29&rft.issn=0162-8828&rft.eissn=1939-3539&rft.volume=39&rft.issue=3&rft.spage=486&rft_id=info:doi/10.1109%2FTPAMI.2016.2552172&rft.externalDBID=NO_FULL_TEXT
thumbnail_l	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=0162-8828&client=summon
thumbnail_m	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=0162-8828&client=summon
thumbnail_s	http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=0162-8828&client=summon