CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation

CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation

ID:40713818

大小:528.63 KB

页数:7页

时间:2019-08-06

CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation _第1页
CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation _第2页
CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation _第3页
CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation _第4页
CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation _第5页
资源描述:

《CVPR2005HuaHuang_Probabilistic Contour Extraction Using Hierarchical Shape Representation 》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、ProbabilisticContourExtractionUsingHierarchicalShapeRepresentation1,2212XinFanChunQiDequnLiangHuaHuang12SchoolofInformationEngineering,SchoolofElectronicsandInformationEngineering,DalianMaritimeUniversity,Xi’anJiaotongUniversity,Dalian,P.R.China.116026Xi’an,P.R.China.710049fanxin@dlmu.e

2、du.cn{qichun,huanghua}@mail.xjtu.edu.cnAbstractarebuiltfromstatisticallyanalyzingexamplesoftheshapesofinterest,namelystatisticalmodels[2,3,17].Inthispaper,weaddresstheissueofextractingRegardlessoftherepresentationofmodels,thecontouroftheobjectwithaspecificshape.Asolutiontotheproblemisfo

3、undbydeterministichierarchicalgraphicalmodelisproposedtorepresentoptimizationtechniquesinmostmodel-basedshapevariations.Acomplexshapeisdecomposedintoapproachestocontourextraction.In[4],DynamicseveralcomponentswhicharedescribedasPrincipalProgrammingisusedtofindagloballyoptimalmatch.Compo

4、nentAnalysis(PCA)basedmodelsinvariousInActiveShapeModels(ASM)proposedbyCootesetlevels.Thehierarchicalrepresentationallowsforal.[2],alocalminimumofthematchingenergyischain-likeconditionaldependencywithinasinglelevelfoundbasedongreedysearchinginlocalandbidirectionalcommunicationbetweendif

5、ferentneighborhoods.Whilethematchingisformulatedaslevels.Additionally,aSequentialMonte-Carlo(SMC)Bayesianestimationin[17],theproblemisresolvedbasedinferencealgorithmthatcanexploretheviaapyramidsearchingstrategy.AndCremersetal.graphicalstructureisproposedtoestimatethecontour.provideagrad

6、ientdescentschemefortheTheexperimentsperformedonreal-worldhandandminimizationofamodifiedMumford-Shahenergyfaceimagesshowthattheproposedmethodiseffectivefunctionalthatincorporatesstatisticalshapeincombatingocclusionandclutteredbackground.information[3].However,itisworthnotingthatthereMor

7、eover,itispossibletoisolatethelocalizationexistmultipleinconsistentinterpretationsofimageerrortoanindividualcomponentofashapeattributedfeaturesduetothecomplicationofnaturalscenes.tothehierarchicalrepresentation.Thus,itmightbeproblematictoperformasearchingprocedurealongonedirect

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。