PatternRecognition43(2010)267--
279
ContentslistsavailableatScienceDirect
PatternRecognition
journalhomepage:/locate/p
r
Anovel3Dmeshcompressionusingmeshsegmentationwithmultipleprincipalplaneanalysis
Shyi-ChyiChenga, ,Chen-TsungKuob,c,Da-ChunWub
a
DepartmentofComputerScienceandEngineering,NationalTaiwanOceanUniversity,2Pei-NingRoad,Keelung202,Taiwan
InstituteofEngineeringScienceandTechnology,NationalKaohsiungFirstUniversityofScienceandTechnology,1UniversityRoad,Yenchao,Kaohsiung824,Taiwanc
DepartmentofInformationManagement,LongcyuanVeteransHospital,VAC,ExecutiveYuan,1AnpingLane1JhaoshengRoad,Pingtung912,Taiwan
b
ARTICLEINFOABSTRACT
Articlehistory:
Received24July2007
Receivedinrevisedform20May2009Accepted26May2009Keywords:3DmeshCompressionSegmentation
Principleplaneanalysisk-meansclustering
Thispaperproposesanovelschemefor3Dmodelcompressionbasedonmeshsegmentationusingmul-tipleprincipalplaneanalysis.Thisalgorithmfirstperformsameshsegmentationscheme,basedonfusionofthewell-knownk-meansclusteringandtheproposedprincipalplaneanalysistoseparatetheinput3Dmeshintoasetofdisjointedpolygonalregions.Theboundaryindexingschemeforthewholeobjectiscreatedbyassemblinglocalregions.Finally,thecurrentworkproposesatriangletraversalschemetoencodetheconnectivityandgeometryinformationsimultaneouslyforeverypatchundertheguidanceoftheboundaryindexingscheme.Simulationresultsdemonstratethattheproposedalgorithmobtainsgoodperformanceintermsofcompressionrateandreconstructionquality.
©2009ElsevierLtd.Allrightsreserved.
1.Introduction
Inrecentyears,three-dimensional(3D)meshmodelshavebeenwidelyusedingeographicaldatabases,manufacturingassemblies,virtualrealityforentertainmentapplications,orotherinteractiveapplications.Thesemodelsareoftenrepresentedascomplextrian-gularmeshes,whichmayhavethousandsorevenmillionsofver-ticesandpolygons.Alloftheseapplicationsrequirelargestorage,computerpowerandaccessoverbandwidth-limitedlinks.Thus,itisessentialtocompressthe3Dmodelsefficiently.SinceDeering[1]firstintroducedtheconceptofgeneralizedtriangularmeshcompres-sion,manyalgorithmscategorizedaslosslessorlossy,single-rateorprogressive,andsingleresolutionormulti-resolution,havebeenproposedtocompress3Dmeshes[1–18].Agoodreviewon3Dmeshcompressiontechnologycanbefoundin[12,13].
Comparedwiththelossy3Dmeshcoding,alosslesstechniquere-constructsmeshdataidenticaltotheoriginal.Manyapplicationsuselosslessdatacompression,suchasinexecutablecode,wordprocess-ingfiles,etc.Tocomparesingle-ratecodingwithprogressivecoding,single-ratecodingfocusesonsavingthebandwidthbetweenCPUandthegraphiccard;however,progressivecompressionof3Dmeshesprovidesmultipleresolutionsfortransmittingcomplexmeshesovernetworkswithlimitedbandwidth.Earlyresearchconducted
Correspondingauthor.
E-mailaddresses:csc@mail.ntou.edu.tw(S.-C.Cheng),dcwu@ccms.nkfust.edu.tw(C.-T.Kuo),jmskuo@mail.vhlc.gov.tw(D.-C.Wu).
0031-3203/$-seefrontmatter©2009ElsevierLtd.Allrightsreserved.doi:10.1016/j.patcog.2009.05.016
single-ratecodingon3Dmeshcompression.Chow[2]presentedanalgorithmtoefficientlyproducegeneralizedtrianglemeshes.HismeshifyingalgorithmsandvariablecompressionmethodachievedcompressionratioshigherthanDeering'smethod[1].Toumaetal.[3]proposedthevalence-drivenapproachthatrecordsthedegreeofeachvertexalongaspiralingvertextree.Progressivemeshcom-pressionhasbeenalsointensivelyresearched,sinceitenablesthedecodertoreconstruct3Dmodelscontinuouslyfromcoarsetofinelevels-of-details(LODs)[4–8].TheEdgebreakermethodproposedbyRossignac[9]canhandlemanifoldmesheswithmultipleboundaryloopsandhandles.Gumholdetal.[10]improvedthisconnectivityupperboundto3.522bpv.
Trianglemeshcompressionhasbeenthefocusofmuchstudy.Thisrepresentationcontainstwokindsofinformation:geometryandconnectivity.Geometrycodingdescribesvertexcoordinatesinthe3Dspace,andconnectivitycodingdescribeshowtoconnecttheseposi-tions.Theconnectivitycompressionproblemhasbeenwellstudied,withmanyexistingmethods[12,13]achievingbit-ratesoflessthantwobitspertrianglefortheconnectivityportionofamesh[11].Lessefforthasgoneintogeometrycompression,oftensimplyperformedbypredictioncodingandquantization[14].Recently,researchershaveproposedothergeometrycompressiontechniques,includingwavelettransform[15],spectralcompression[16],andk–dtreeoroctreedecomposition[17].Althoughtheperformanceofthese3Dmeshcompressiontechniquesachievingbit-ratesbetween1and2bytespervertexisgood,theoutputbitstreamremainslargeforcomplexobjectswithhighnumbersofverticesandtriangles.Thus,developingeffectivecompressiontechniquesforthevertexdatato
搜索“diyifanwen.net”或“第一范文网”即可找到本站免费阅读全部范文。收藏本站方便下次阅读,第一范文网,提供最新人文社科A novel 3D mesh compression using mesh segmentation with multiple principal plane analysis全文阅读和word下载服务。
相关推荐: