基于局部信息统计的人耳识别方法.pdfVIP

ComputerEngineeringandApplications计算机工程与应用 @图形、图像、模式识别@ 基于局部信息统计的人耳识别方法 王晓云1，郭金玉2 WANG Xiaoyunl。GUOJinyu2 1．沈阳理工大学机械工程学院，沈阳110168 10142 2．沈阳化工大学信息工程学院，沈阳l 1．SchoolofMechanical 1101 Engineering，ShenyangLigongUniversity,Shenyang68，China 2．SchoolofInformation ofChemical 110142，China Engineering，ShenyangUniversity Technology,Shenyang WANG ear basedonstatisticfeaturesoflocal Xiaoyun，GUOJinyu．Humanrecognition information．Computer and Applications，2012，48(9)：141—144． newhumanear onstatisticfeaturesoflocal humanear Abstract：A recognitionapproach，based information，isproposed．A gray‘scale isdividedintoseveral assortedfeaturesofeach areabstracted．Featureof are in- image sub—regions．The sub—regions sub—regionsjoined a tobuildhumanfeature a aboutthes仃llcnlreandlocalinformation to featurevector ear vector,thus givescomprehensivedescription ex- ofhumanear classificationis thenearest classifier．Threedifferentfeature images．Thepattem implementedbyapplying neighbor tractionmethodsare andUSTBhumaneardatabaseare to this resultsshowthat adopted appliedtestingalgorithm．Theexperimental the rateis morethan with informationandthemethodbasedonstatisticsfbanlreoflocalin— recognitionimproved 30％comparedglobal formationis effective． proved words：ear transform Key recognition；localinformation；statisticfeatures；Fourier 摘要：提出了一种基于局部信息统计的人耳识别方法。该方法将一幅人耳图像分成若干个子区域，分别提取每个子区域的分 类特征，将各个子区域的特征串联为一个特征向量构筑人耳特征矢量，更加全面描述了人耳图像的局部与结构信