资源描述：
《边缘检测微分算子的分析及在医学图像中的应用》由会员上传分享，免费在线阅读，更多相关内容在工程资料-天天文库。
1、Seediscussions,stats,andauthorprofilesforthispublicationat:https://www.researchgate.net/publication/7972542ThefrequencyfeaturesandapplicationofedgedetectiondifferentialoperatorsinmedicalimageArticleinShengwuyixuegongchengxuezazhi=Journalofbiomedicalengineering=Shengwuyixuegongchengxuezazhi·Mar
2、ch2005Source:PubMedCITATIONSREADS4295authors,including:JianWuHuiDingTsinghuaUniversityTsinghuaUniversity42PUBLICATIONS257CITATIONS69PUBLICATIONS178CITATIONSSEEPROFILESEEPROFILEGuangzhiWangTsinghuaUniversity246PUBLICATIONS714CITATIONSSEEPROFILESomeoftheauthorsofthispublicationarealsoworkingonth
3、eserelatedprojects:MedicalRoboticsforInterventionalSurgeryViewprojectRehabilitationEngineeringViewprojectAllcontentfollowingthispagewasuploadedbyGuangzhiWangon04January2017.Theuserhasrequestedenhancementofthedownloadedfile.生物医学工程学杂志2005;22(1)∶82～85JBiomedEng边缘检测微分算子的分析及在医学图像中的应用$山吴 剑 丁 辉 王广志 丁
4、海曙 周伊豆页(清华大学医学院生物医学工程系,100084) 摘要 边缘检测是医学图像处理必不可少的一步,采用微分算子检测边缘是最常用也是处理效果比较好的一种方法。我们从频域角度出发,分析了Roberts算子、Prewitt算子、Sobel算子以及拉普拉斯算子的频谱特性。提出了在使用和构造微分算子时,要充分考虑它们的频谱特性。不同的微分算子对不同边缘检测的敏感程度是不同的,因此对不同类型的边缘提取,应该采用对此类型边缘敏感的算子进行边缘提取。最后通过实际的MRI图像处理验证了边缘检测算子选择的重要性。关键词 边缘检测 微分算子 医学图像TheFrequencyFeaturesan
5、dApplicationofEdgeDetectionDifferentialOperatorsinMedicalImage$WuJianDingHuiWangGuangzhiDingHaishuZhouYiyi(DepartmentofBiomedicalEngineering,SchoolofMedicine,TsinghuaUniversity,Beijing100084,China)AbstractEdgedetectionisanabsolutelynecessarystepinmedicalimageprocessing,andtheuseofdifferentialo
6、peratorstodetectedgeisoneofthemostcommonandeffectivemethods.InthispaperareanalyzedthefrequencyfeaturesoftheRobertsoperator,Prewittoperator,SobeloperatorandLaplacianoperatorfromtheviewpointoffrequencydomain,anditisproposedthatthefrequencyfeaturesofthedifferentialoperatorsshouldbeconsideredwhend
7、ifferentialoperatorisbeingusedandöorconstructed.Becauseedgedetectionoperatorissensitivetotheedgetype,theappropriateoperatorshouldbeadoptedindifferentedgetypedetection.Finally,theimportanceandnecessityofselectingedgedetectionoperatorarev
