Region-Based Hierarchical Image Matching

Region-Based Hierarchical Image Matching

ID:40091372

大小:1.37 MB

页数:20页

时间:2019-07-20

Region-Based Hierarchical Image Matching_第1页
Region-Based Hierarchical Image Matching_第2页
Region-Based Hierarchical Image Matching_第3页
Region-Based Hierarchical Image Matching_第4页
Region-Based Hierarchical Image Matching_第5页
资源描述:

《Region-Based Hierarchical Image Matching》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库

1、IntJComputVisDOI10.1007/s11263-007-0077-5Region-BasedHierarchicalImageMatchingSinisaTodorovic·NarendraAhujaReceived:16September2006/Accepted:17July2007©SpringerScience+BusinessMedia,LLC2007AbstractThispaperpresentsanapproachtoregion-basedisomorphism·Segmentationtr

2、ees·Transitiveclosures·hierarchicalimagematching,where,giventwoimages,theAssociationgraphs·Maximumweightcliquesgoalistoidentifythelargestpartinimage1anditsmatchinimage2havingthemaximumsimilaritymeasuredefined1Introductionintermsofgeometricandphotometricpropertiesof

3、regions(e.g.,area,boundaryshape,andcolor),aswellasregionImagematchingisalong-standingproblemincomputervi-topology(e.g.,recursiveembeddingofregions).Tothisend,sion.Whilemostapproachesusepointfeaturesand/orcurveeachimageisrepresentedbyatreeofrecursivelyembed-fragmen

4、tsforimagematching,thereisalsoasignificantdedregions,obtainedbyamultiscalesegmentationalgo-amountofworkonregion-basedmatching.Forexample,rithm.Thisallowsustoposeimagematchingasthetreematchingusingregionpropertiesisdonetoaddressprob-matchingproblem.Toovercomeimaging

5、noise,one-to-one,lemsfromstereomatching(e.g.,MedioniandNevatia1985;many-to-one,andmany-to-manynodecorrespondencesareCohenetal.1989a;RandriamasyandGagalowicz1991),allowed.Thetreesarefirstaugmentedwithnewnodesgen-andmotion/optical-flowanalysis(e.g.,XuguangandRam-erate

6、dbymergingadjacentsiblingnodes,whichproduceschandran1999;Ming-Hsuanetal.2002;FuhandMaragosdirectedacyclicgraphs(DAGs).Then,transitiveclosuresof1989)toobjectrecognition(e.g.,BasriandJacobs1997;theDAGsareconstructed,andthetreematchingproblemKeselmanandDickinson2005)

7、.Thisworkdemonstratesreformulatedasfindingabijectionbetweenthetwotran-severaladvantagesofusingregionsoverinterestpointsorsitiveclosuresonDAGs,whilepreservingtheconnectivityedgefragmentsforimagematching.Thehigherdimension-andancestor-descendantrelationshipsoftheorig

8、inaltrees.alityofregionsmakesthemricherdescriptorsoftargetob-Theproposedapproachisvalidatedonrealimagesshow-jects’geometricproperties,suchassizeandshape

当前文档最多预览五页,下载文档查看全文

此文档下载收益归作者所有

当前文档最多预览五页,下载文档查看全文
温馨提示:
1. 部分包含数学公式或PPT动画的文件,查看预览时可能会显示错乱或异常,文件下载后无此问题,请放心下载。
2. 本文档由用户上传,版权归属用户,天天文库负责整理代发布。如果您对本文档版权有争议请及时联系客服。
3. 下载前请仔细阅读文档内容,确认文档内容符合您的需求后进行下载,若出现内容与标题不符可向本站投诉处理。
4. 下载文档时可能由于网络波动等原因无法下载或下载错误,付费完成后未能成功下载的用户请联系客服处理。