资源描述：
《Statistical shape models for 3D medical image segmentation, A review (1)》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、MedicalImageAnalysis13(2009)543–563ContentslistsavailableatScienceDirectMedicalImageAnalysisjournalhomepage:www.elsevier.com/locate/mediaStatisticalshapemodelsfor3Dmedicalimagesegmentation:AreviewTobiasHeimann*,Hans-PeterMeinzerDiv.ofMedicalandBiologicalInformati
2、cs,GermanCancerResearchCenter,ImNeuenheimerFeld280,D-69120Heidelberg,GermanyarticleinfoabstractArticlehistory:Statisticalshapemodels(SSMs)havebynowbeenfirmlyestablishedasarobusttoolforsegmentationofReceived27August2007medicalimages.While2Dmodelshavebeeninusesincet
3、heearly1990s,wide-spreadutilizationofReceivedinrevisedform24February2009three-dimensionalmodelsappearedonlyinrecentyears,primarilymadepossiblebybreakthroughsinAccepted18May2009automaticdetectionofshapecorrespondences.Inthisarticle,wereviewthetechniquesrequiredtoc
4、re-Availableonline27May2009ateandemploythese3DSSMs.Whileweconcentrateonlandmark-basedshaperepresentationsandthoroughlyexaminethemostpopularvariantsofActiveShapeandActiveAppearancemodels,wealsoKeywords:describeseveralalternativeapproachestostatisticalshapemodeling
5、.StructuredintothetopicsofshapeStatisticalshapemodelrepresentation,modelconstruction,shapecorrespondence,localappearancemodelsandsearchalgo-DeformablesurfaceActiveShapemodelrithms,wepresentanoverviewofthecurrentstateoftheartinthefield.WeconcludewithasurveyActiveAp
6、pearancemodelofapplicationsinthemedicalfieldandadiscussionoffuturedevelopments.Ó2009ElsevierB.V.Allrightsreserved.1.Introductionconcentrateonmethodsforvolumetricimages.However,manymodelingmethodshaveonlybeenappliedto2Dsofar.TotheInthelasttwodecades,model-basedsegm
7、entationapproachesextentthatsuchmethodscanbegeneralizedorextendedtothehavebeenestablishedasoneofthemostsuccessfulmethodsfor3Dcase,wehaveincludedtheminthepresentreview.Wehaveimageanalysis.Bymatchingamodelwhichcontainsinformationfurthermoreincludedsomemethodswhichh
8、aveshowntobeveryabouttheexpectedshapeandappearanceofthestructureofinter-successfulin2D,butwhicharetechnicallynotfeasiblein3D–esttonewimages,thesegmentationisco
