简介：Thispaperpresentsanewclassofsurfacesthatgivetwoquitedifferentappearanceswhentheyareseenfromtwospecialviewpoints.Theinconsistentappearancescanbeperceivedbysimultaneouslyviewingthemdirectlyandinamirror.Thisphenomenonisanewtypeofopticalillusion,andwehavenameditthe"ambiguouscylinderillusion",becauseitistypicallygeneratedbycylindricalsurfaces.Weconsiderwhythisillusionarises,andwepresentamathematicalmethodfordesigningambiguouscylinders.

简介：Thispaperpresentsanewalgorithmforgenerating3DimagesofB-repsobjectswithtrimmedsurfaceboundaries.The3Dimageisadiscretevoxel-maprepresentationwithinaCubicFrameBuffer(CFB).Thedefinitionof3Dimagesforcurve,surfaceandsolidobjectareintroducedwhichimplytheconnectivityandfidelityrequirements.AdaptiveForwardDifferencingmatrix(AFD-matrix)for1D-3Dmanifoldsin3Dspaceisdeveloped.BysettingrulestoupdatetheAFD-matrix,theforwarddifferencedirectionandstepwisecanbeadjusted.Finally,anefficientalgorithmispresentedbasedontheAFD-matrixconceptforconvertingtheobjectin3Dspaceto3Dimagein3Ddiscretespace.

简介：Inrecentyears,accordingtotheneedofintelligentvideosurveillancesystemincreasingrapidlyinmetropolitancities,adesignbasedonS3C2440microprocessorandembeddedLinuxoperatingsystemisadoptedforreal-timevideotargettracking.However,itisverychallengingasembeddedsystemsusuallyaffordlimitedprocessingpowerandlimitedresources.Therefore,toaddressthisproblem,areal-timetrackingalgorithmusingmulti-featuresbasedoncompressivesensingisproposedandimplemented.Thealgorithmusesmultiplematrixastheprojectionmatrixofthecompressivesensingandthecompresseddateasthemultiplefeaturestoextractusefulinformationneededbytrackingprocess.FunctionsandlibrariesinOpenCVwhichweredevelopedbyIntelCorporationareutilizedforbuildingthetrackingalgorithms.Itistestedwithvariantvideosequencesandtheresultsshowthatthealgorithmachievesstabletrackingforthetargetmovedofthelightchanged.

简介：

简介：

简介：3Dobjectscanbestoredincomputerofdifferentdescribingways,suchaspointset,polyline,polygonalsurfaceandEuclideandistancemap.Momentinvariantsofdifferentordersmayhavethedifferentmagnitude.Amethodfornormalizingmomentsof3Dobjectsisproposed,whichcansetthevaluesofmomentsofdifferentordersroughlyinthesamerangeandbeappliedtodifferent3Ddataformatsuniversally.Thenaccuratecomputationofmomentsforseveralobjectsispresentedandexperimentsshowthatthiskindofnormalizationisveryusefulformomentinvariantsin3Dobjectsanalysisandrecognition.

简介：