3D Mean-Shift Tracking of Human Body Parts

3D Mean-Shift Tracking of Human Body Parts

ID:39908640

大小:952.41 KB

页数:12页

时间:2019-07-14

3D Mean-Shift Tracking of Human Body Parts_第1页
3D Mean-Shift Tracking of Human Body Parts_第2页
3D Mean-Shift Tracking of Human Body Parts_第3页
3D Mean-Shift Tracking of Human Body Parts_第4页
3D Mean-Shift Tracking of Human Body Parts_第5页
资源描述:

《3D Mean-Shift Tracking of Human Body Parts》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、3DMean-ShiftTrackingofHumanBodyPartsandRecognitionofWorkingActionsinanIndustrialEnvironmentMarkusHahn1,FuadQuronfuleh1,ChristianW¨ohler2,andFranzKummert31DaimlerAG,GroupResearchandAdvancedEngineering,Ulm,Germany2ImageAnalysisGroup,DortmundUniversityofTechnology,Dortmund,Germany3Applied

2、Informatics,BielefeldUniversity,Bielefeld,GermanyAbstract.Inthisstudywedescribeamethodfor3Dtrajectorybasedrecognitionofanddiscriminationbetweendifferentworkingactionsinanindustrialenvironment.Amotion-attributed3Dpointcloudrepresentsthescenebasedonimagesofasmall-baselinetrinocularcameras

3、ystem.Atwo-stagemean-shiftalgorithmisusedfordetectionand3Dtrack-ingofallmovingobjectsinthescene.Asequenceofworkingactionsisrecognisedwithaparticlefilterbasedmatchingofanon-stationaryHiddenMarkovModel,relyingonspatialcontextandaclassificationoftheobserved3Dtrajectories.Thesystemisabletoex

4、tractanobjectperformingaknownactionoutofamultitudeoftrackedobjects.The3Dtrackingstageisevaluatedwithrespecttoitsmetricaccuracybasedonninereal-worldtestimagesequencesforwhichgroundtruthdataweredetermined.Anexperimentalevaluationoftheactionrecognitionstageisconductedusing20real-worldtest

5、sequencesacquiredfromdif-ferentviewpointsinanindustrialworkingenvironment.Weshowthatoursystemisabletoperform3Dtrackingofhumanbodypartsandasubsequentrecognitionofworkingactionsunderdifficult,realisticcon-ditions.Itdetectsinterruptionsofthesequenceofworkingactionsbyenteringasafetymodeandre

6、turnstotheregularmodeassoonastheworkingactionscontinue.1IntroductionTheefficiencyofmanyindustrialproductionprocessescanbeincreasedbyes-tablishingaclosecollaborationbetweenhumansandmachinesexploitingtheiruniquecapabilities.Asafeinteractionbetweenhumansandindustrialrobotsrequiresvisionmeth

7、odsfor3Dposeestimation,tracking,andrecognitionofthemotionofhumanbodyparts.Arobustmethodoftenusedforobjecttrackingisthemean-shiftalgorithm[2,3,5],whichsearchesforalocalmodeoftheempiricaldensityfunction.Themean-shifttrackingapproachintroducedin[4]isextendedto3Dspacein[12].Intheexperime

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。