资源描述：
《一种各向异性wells算法脑核磁共振图像分割模型》由会员上传分享，免费在线阅读，更多相关内容在教育资源-天天文库。
1、计算机研究与发展ISSN10001239CN111777TPJournalofComputerResearchandDevelopment47(11):18781885,2010一种各向异性Wells算法脑核磁共振图像分割模型1223陈允杰张建伟王顺凤詹天明1(南京信息工程大学数理学院南京210044)2(南京信息工程大学滨江学院南京210044)3(南京理工大学计算机科学与技术学院南京210094)(generalcyj@yahoo.com.cn)BrainMRImageSegmentationBasedonA
2、nisotropicWellsModel1223ChenYunjie,ZhangJianwei,WangShunfeng,andZhanTianming1(SchoolofMathandPhysics,NanjingUniversityofInformationScienceandTechnology,Nanjing210044)2(BinjiangSchool,NanjingUniversityofInformationScienceandTechnology,Nanjing210044)3(SchoolofComputerScience
3、andTechnology,NanjingUniversityofScienceandTechnology,Nanjing210094)AbstractNuclearmagneticresonance(MR)imageanalysishasbecomeamajormeansoftheauxiliarymedicalservices.However,intensityinhomogeneity,whichisusuallynamedasbiasfield,causesconsiderabledifficultyinthequantitati
4、veanalysisofMRimages.ThusbiasfieldestimationisanecessarypreprocessingstepbeforequantitativeanalysisofMRdata.TheWellsmodel,oneofthewidelyusedmethods,usesknowledgeoftissueintensitypropertiesandintensityinhomogeneitiestocorrectandsegmentMRimages.However,theclassicalWellsmode
5、lonlyusestheintensityinformationandnospatialinformationistakenintoaccount,soitissensitivetothenoise.Inordertoovercomethislimitation,theGibbstheoryandtheimagestructureinformationareusedtoconstructanisotropicGibbsrandomfield.ThetraditionalGibbstheoryusuallylosestheinformatio
6、nofthebeamstructureregionsandthecornerregions.Withthespatialinformation,theanisotropicGibbsrandomfieldcanreducetheeffectofthenoiseandcontaintheinformationofthebeamstructureregionsandthecornerregions.TheanisotropicGibbsrandomfieldisincorporatedintotheWellsmodel.Theexperimen
7、tsofsegmentingthebrainmagneticresonanceimagesshowthattheproposedmethodcanobtainbetterresultsinanaccurateway.KeywordsWellsmodel;Gibbsrandomfield;anisotropicGibbsrandomfield;imagesegmentation;structuretensor摘要核磁共振图像分析已经成为主要的医疗辅助手段之一.然而,由于偏移场的影响导致该类图像的分析较为困难,去偏移场已成为图像分析的必要
8、步骤.Wells算法将图像分割和去偏移场放入同一个框架内并取得较好的结果.然而该算法没有考虑像素间的位置信息,因而导致该算法对图像噪声比较敏感.为了克服其局限性,利用Gibbs理
