(李鹏)核磁共振图像MRI去噪方法研究与实现--论文-(不涉及人名)

武汉科技大学本科毕业论文本科毕业论文题目：核磁共振图像MRI去噪方法研究与实现学院:计算机科学与技术学院专业:计算机科学与技术专业学号:20071313****学生姓名:****指导教师:*****日期:2011年6月9日武汉科技大学本科毕业论文I摘要磁共振成像(nonnImnI)已经成为临床诊断必不可少的手段，但是快速成像以及高分辨率MRI使重构后的磁共振（MR）图像呈现出较大的噪声伪影，影响图像的质量，从而影响后续的临床诊断和治疗。由于用于视觉检测的幅度图像中的噪声并不是加性高斯白噪声，而是呈现出与信号相关n分布，使得传统的针对高斯分布噪声的方法无法很好地去除MRI图像的噪声。而这些噪声伪影都是来自原始K-空间复数数据实部和虚部不相关的零均。我们于是提出了一种新的迭代非局部均值滤波的图像去噪方法.权系数的计算依赖每次迭代更新得到的图像,同时对迭代更新后得到的图像进行加权平均.这样就避免了权系数的计算以及加权平均所用的图像的不一致所带来的图像边缘模糊以及对比度不清晰的现象.还证明了新的迭代方法满足极大极小原则.实验结果表明,该方法去噪的同时能较好地保持图像的边缘以及细小结构。关键字：磁共振成像；图像去噪；非局部滤波；加权平均；扩散方程武汉科技大学本科毕业论文IIAbstractMagneticResonancehanging(MRI)playsaninreplaceableroleinclinicdiagnosis，However,fastimagingorhighresolutionachievementoftenleadtogreatnoiseartifactsinreconstructedmagneticresonance(MR)images,whichinterferewithsubsequentclinicaldiagnosisandcure.ThenoiseinMRmagnitudeimagesobeysaRiciandistribution.Riciannoise,unlikeadditiveGaussiannoise,issignal-dependentandconsequentlynoiseremovalisadifficulttask.Asweknow,thenoisecontributions,arisingfromtherealandimaginarypartsofthek-spacecomplexdata,areadditive,assumedtobeuncorrelated,andcharacterizedbyazero-meanGaussiannoise.So,itisavailabletodenoisetherealandimaginarypartsofthecompleximages.Aniterativenon-localmeansfilterforimagedenoisingisproposed.Fortheproposedmethod,thecomputationofweightsdependsontheupdatedimage,andforeachiteration,theweightedaveragingisperformedovertheupdatedimage.Thisavoidstheproblemsthattheedgeoftheimageisblurryandthattheimageisnotclear,whicharecausedbythedisagreementofthecomputationoftheweightcoefficientsandthecomputationoftheaveraging.Themax-minprincipleisalsoguaranteedfortheproposediterativenon-localmeansfilter.Numericalexamplesillustratethattheproposedmethodremovesnoisewellwhilepreservingimageedgesandfinedetails.Keywords:magneticresonanceimaging;ImageDenoising;non-localmeansfilter;weightedaveraging;diffusionequation；武汉科技大学本科毕业论文III目录1绪论..................................................................................................................11.1选题研究的背景........................................................................................11.1.1MRI的发展历史...................................................................................11.1.2MRI的形成...........................................................................................21.2MRI在医学临床上的应用.........................................................................31.3磁共振图像的噪音....................................................................................41.3.1噪音的来源...........................................................................................41.3.2MRI图像的噪音模型............................................................................51.4完成的工作................................................................................................62脑核磁共振图像MRI的处理...........................................................................72.1MRI图像噪音............................................................................................72.1.1噪音分类...............................................................................................72.1.2噪音定义...............................................................................................72.1.3性能评价...............................................................................................82.2非局部平滑滤波（NON-LOCALMEANS）.....................................................102.2.1NL-means算法中的方法噪音...........................................................102.2.2NL-means算法一致性.......................................................................122.2.3NL–means算法的设计思路..............................................................132.2.4NL–means算法相关代码..................................................................153软件分析与介绍.............................................................................................183.1GDI+介绍.................................................................................................183.2软件介绍.................................................................................................183.3试验结果及分析......................................................................................213.3.1图像加噪.............................................................................................213.3.2图像滤波.............................................................................................224NL–MEANS算法与其他去噪算法的比较.....................................................27武汉科技大学本科毕业论文IV4.1中值滤波.................................................................................................274.1.1中值滤波简介.....................................................................................274.1.2中值滤波实现原理..............................................................................274.1.3中值滤波核心代码..............................................................................274.2低通滤波.................................................................................................294.2.1低通滤波简介..................................................................................