您好,欢迎访问三七文档
摘要我们常说的生物特征识别技术,其实就是将计算技术应用到其中,然后将生物体的某些具体的行为特性以及一些固有的生理特性,与之前建立的数据库之间的数据进行对比、参照,进而识别某人或者某生物体身份的技术。近年来,有两大热门研究方向,也就是:人眼定位和人脸识别。本篇论文的主要内容是:在人脸识别的基础上,对眼部进行识别标注。首先,对用FLDA(Fliser线性判别分析法),2DPCA(二维主成分分析法),以及PCA(主成分分析法)做成的三种人脸识别方法进行一个简单的说明和比较。然后,综合讨论和分析在PTL(人眼定位算法)中所涉及到的噪声敏感度和人眼定位所需要耗费的时间。具体点说,就是先对样本的人脸图片进行Cabor小波变换,以找到其中的眉眼所在区域。随之,将直接获得低型地形特点的技术应用于眉眼区域,然后我们找到大概的人眼所在区域,利用相似性和对称性找到具体的人眼所在区。最终,要对这个样本人脸图片进行加噪,然后重复上述操作,最后通过观察有无噪声的人眼定位效果,以得出此方法的有效性。次之,介绍本论文的最核心部分:由粗到精的人眼定位法,其主要采用了方差滤波和地形特征点检验两种技术。第一,利用地形特征点检验技术和Cabor小波变换技术,直接获得人眼大致所在区域,也就是所谓的粗定位。第二,确定人脸所在区域,然后再这个区域上利用差分滤波器进一步确定人眼所在位置,也就是所谓的精定位。Bvertex关键字:生物特征;人眼识别;Cabor小波变换;人脸识别;由粗到精的人眼定位ABSTRACTInfact,weoftensaybiometrictechnologythatthecomputingtechnologywasusedintoit,andthedatabetweenspecificbehavioralcharacteristicsoforganismsaswellassomeoftheinherentphysicalcharacteristics,andpriortotheestablishmentofadatabasecomparisonandreference,andtoidentifythepersonorabiometricidentitytechnology.Inrecentyears,therearetwopopularresearchdirections,thatare:eyelocationandfacerecognition.Themaincontentsofthispaperare:onthebasisoffacerecognition,identificationmarkontheeye.First,TheFLDA(Fliserlineardiscriminantanalysis),The2DPCA(dimensionalprincipalcomponentanalysis),andThePCA(PrincipalComponentAnalysis)conductedthreefacerecognitionmethodthatmadeasimpledescriptionandcomparison.Then,acomprehensivediscussionandanalysis,inThePTL(eyelocationalgorithm),isinvolvedinthenoisesensitivityofthehumaneyepositioningtime-consumingneed.Specifically,itisthefirstfaceofthesampleimagetoCabortransform,tofindthelocationwherethefacialfeatures.Followingthis,thetypeofterrainfeaturesdirectaccesstolowtechnologiesinfacialarea,andthenwefindanapproximatelocationofthehumaneye,theuseofsimilarityandsymmetryofthedistricttofindaspecificpersoneye.Ultimately,wewanttofacethissampleimagetoaddnoise,andthenrepeattheoperation,andfinallywekownwhetherthenoiseeffecteyelocationbyobserving,inordertoobtaintheeffectivenessofthismethod.Last,themostcentralpartofthispaper:Themethodtofindeyeslocationfromcoarse,themainuseofthevariancefilteringandterrainfeaturepointtesttwotechnologies.First,theuseofterrainfeaturespointinspectiontechniquesandCaborwavelettransform,directaccesstothehumaneyegenerallyArea,whichistheso-calledcoarsepositioning.Second,weneedtodeterminethelocationofface,thenthisareaisdeterminedusingdifferentialfilterfurthereyelocation,alsoknownasfinepositioning.Keywords:Biometrics;Recognitionofthehumaneye;Caborwavelettransform;facerecognition;fromcoarsetofineeyelocation目录第1章引言..............................................................51.1简介..............................................................51.1.1面部识别技术.................................................51.1.2眼部识别技术.................................................61.2研究背景和意义....................................................81.3论文结构和研究内容................................................81.4小结..............................................................9第2章若干人脸识别方法的比较研究.......................................102.1基于2DPCA和PCA的人脸识别系统...................................102.1.1主成分分析(PCA)算法.......................................102.1.2奇异值分解(SVD)技术.......................................112.1.3二维主成分分析法(2DPCA)技术...............................132.2基于Fliser线性判别分析法(FLDA)的人脸识别方法..................162.3进行分类器设计...................................................192.4仿真实验.........................................................202.5小结.............................................................22第3章基于地形特征和Cabor小波变换的人眼定位算法.......................233.1准备工作.........................................................243.1.1Cabor小波变换...............................................243.1.2形态学算子之“膨胀”.......................................253.1.3灰度积分投影...............................................263.1.4地形特征...................................................263.2基于直接提取地形特征和Cabor小波变换的人眼定位算法...............273.2.1以Cabor小波变换为基础的人脸眉眼区域表示...................273.2.2以直接提取地形特征为基础的的人眼候选点检测.................283.2.3人眼定位之对称相似度分析法.................................303.3仿真实验.........................................................313.3.1在没有噪声的情况下的仿真实验...............................313.3.2在有噪声的情况下的仿真实验.................................333.4小结.............................................................34第4章一种由粗到精的人眼定位算法.......................................354.1人眼精确定位之人眼方差滤波器.....................................354.1.1人眼方差滤波器.............................................354.1.2进行人眼精确定位............................................364.2眼部识别实现的基本流程图.........................................394.3实验结果以及结果分析.............................................404.3.1在没有噪声的情况下,PTL和GPL的仿真效果对比实验............404.3.2在有噪声的情况下,PTL和GPL的仿真效果对比实验..............414.3.3眼部识别的效果及其分析......................................424.4小结.............................................................46结论....................................................................48致谢....................................................................50参考文献....................
本文标题:眼部识别(毕设)
链接地址:https://www.777doc.com/doc-4224661 .html