资源描述:
《基于二维经验模态分解的医学图像融合算法》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、ISSN1000-9825,CODENRUXUEWE-mail:jos@iscas.ac.cnJournalofSoftware,Vol.20,No.5,May2009,pp.1096−1105http://www.jos.org.cndoi:10.3724/SP.J.1001.2009.03542Tel/Fax:+86-10-62562563©byInstituteofSoftware,theChineseAcademyofSciences.Allrightsreserved.∗基于二维经验模态分解的医学图像融合算法1+1,2郑有志,覃征1(
2、清华大学计算机科学与技术系,北京100084)2(清华大学软件学院,北京100084)MedicalImageFusionAlgorithmBasedonBidimensionalEmpiricalModeDecomposition1+1,2ZHENGYou-Zhi,QINZheng1(DepartmentofComputerScienceandTechnology,TsinghuaUniversity,Beijing100084,China)2(SchoolofSoftware,TsinghuaUniversity,Beijing100084
3、,China)+Correspondingauthor:E-mail:zhengyz04@mails.tsinghua.edu.cnZhengYZ,QinZ.Medicalimagefusionalgorithmbasedonbidimensionalempiricalmodedecomposition.JournalofSoftware,2009,20(5):1096−1105.http://www.jos.org.cn/1000-9825/3542.htmAbstract:Anadaptivemedicalimagefusionalgori
4、thmbasedontherepresentationofbidimensionalempiricalmodedecomposition(BEMD)isproposed.Sourcemedicalimagesaredecomposedintoanumberofbidimensionalintrinsicmodefunctions(BIMF)aswellasaresidualimage.ImagefeaturesareextractedthroughHilbert-HuangtransformontheBIMF.ThenthecompositeB
5、EMDisformedbyregion-basedfusionrulesondatarepresentationsofBEMD.Finally,thefusedimageisobtainedbyinverseBEMDonthecompositerepresentation.TheBEMDisanadaptivedatadecompositionrepresentation,andhasbetterperformancethanFourierandwavelettransform.Theproposedalgorithmdoesnotneedpr
6、edeterminedfiltersorwaveletfunctions.Experimentalresultsshowthattheproposedalgorithmprovidessuperiorperformanceoverconventionalfusionalgorithmsinimprovingthequalityoffusedimages.Keywords:imagefusion;bidimensionalempiricalmodedecomposition(BEMD);Hilbert-Huangtransform;region-
7、basedfusionrule;regionsegmentation摘要:提出了一种自适应的二维经验模态分解(bidimensionalempiricalmodedecomposition,简称BEMD)医学图像融合算法.待融合的医学图像经过BEMD分解成二维的内蕴模函数(bidimensionalintrinsicmodefunction,简称BIMF)和趋势图像.BIMF图像经过Hilbert-Huang变换提取图像特征,然后,图像分解的各部分数据在区域融合规则下形成综合BEMD表示.最后,综合BEMD表示进行BEMD逆变换得到融合后的医学
8、图像.BEMD分解方法是一种完全自适应的数据分解表达形式,具有比Fourier变化和小波分解更好的特性.该医学图像融合算法不需要预先定义滤波器或小波函