Fitting a Morphable Model to Pose and Shape of a Point Cloud

Fitting a Morphable Model to Pose and Shape of a Point Cloud

ID:39221187

大小:3.31 MB

页数:8页

时间:2019-06-27

Fitting a Morphable Model to Pose and Shape of a Point Cloud_第1页
Fitting a Morphable Model to Pose and Shape of a Point Cloud_第2页
Fitting a Morphable Model to Pose and Shape of a Point Cloud_第3页
Fitting a Morphable Model to Pose and Shape of a Point Cloud_第4页
Fitting a Morphable Model to Pose and Shape of a Point Cloud_第5页
资源描述:

《Fitting a Morphable Model to Pose and Shape of a Point Cloud》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、FittingaMorphableModeltoPoseandShapeofaPointCloudDavidC.Schneider,PeterEisertFraunhoferHeinrichHertzInstitute,Einsteinufer37,10587Berlin,Germany{david.schneider,eisert}@hhi.fraunhofer.deAbstractmodelshavebeenusedmostsuccessfullyinthedo-mainofthehumanheadan

2、dface,wheretheyareThepaperaddressesthetaskoffittingamorphablecalledmorphableheadmodels[9].Awidenumberheadmodeltoadense,unstructuredanduntexturedofapplicationsincomputervisionandgraphicscanpointcloud.Theproblemistypicallyapproachedbedrivenorsupportedbymorpha

3、blemodels,in-inamulti-stepprocess,comprisingofgenericnon-cluding,forexample,facerecognition[3,10],facerigidregistration,conversiontothemodel’stopol-tracking[17,18]andseveralvarietiesof3Drecon-ogyandfittingofthemodel.Here,adirectapproachstruction,e.g.fromunc

4、alibratedvideo[12]orfromisproposedwherethemorphablemodelisfittedstereoimages[2].tothepointclouditselfinanoptimizationprocessInordertofitthemorphablemodeltotheshapefollowingtheIterativeClosestPointsframework.ofahumanhead—i.e.tocomputethemodelparam-Alongwithsh

5、apeparameters,rigidposeandscaleetersthatdescribethehead—thetargetheadmustareestimatedexplicitlywhichleadstobetterfittingberepresentedinawaythatissemanticallyconsis-resultsthanshapeestimationalone.Acompactfor-tentwiththemodel:Thetargetmusthavethesamemulation

6、ofacostfunctionisproposed,applicablenumberofverticesasthemodelandtheseverticesalsointhecaseofamodelcomprisingofmultiple,mustbeplacedatsemanticallyequivalentlocations.independentsub-modelseachofwhichdescribesaIf,forexample,thevertexwithindexKdescribesregion

7、oftheface.Theuseofthealgorithmforthetipofthenoseinthemodel,thismustalsobeothertaskssuchasaddingdepthtomugshot-stylethecaseintherepresentationofthetargethead.facephotosisdiscussed.Notethatsemanticconsistencyisnotthesameas(anddoesnotrequire)topologicalequiva

8、lenceas1Introductiontopologydoesnotrelatetosemanticsandmattersoftriangulationarenotimportantforamorphable“Morphablemodels”areadaptivemodelsof3Dmodel.shape(andsometimestexture)ofacertainobjectInpracticethiscon

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。