局部结构保持的鉴别分析方法  

DISCRIMINANT AND ANALYSE METHOD OF LOCAL STRUCTURE PRESERVING

作　　者：王颖静[1] 王正群[1] 张国庆[1] 俞振洲[1] 

机构地区：[1]扬州大学信息工程学院,江苏扬州225009

出　　处：《计算机应用与软件》2011年第12期51-53,118,共4页Computer Applications and Software

基　　金：国家自然科学基金(60875004);江苏省自然科学基金项目(BK2009184);江苏省高校自然科学基金项目(07KJB520133)

摘　　要：局部投影保持LPP(Locality Preserving Projections)是一种局部特征提取算法,它能够有效地保留数据集的局部结构。不相关保局投影鉴别UDLPP(Uncorrelated Discriminant Locality Preserving Projections)在LPP的基础上考虑了类别信息,通过保留类内几何结构并最大化类间距离获得了良好的鉴别性能。结合UDLPP的思想,在UDLPP的基础上提出了一种局部结构保持的鉴别分析方法PCLSP(Pattern Classification based on Local Structure Preserving)。该方法结合了数据集的类别信息以及数据集的局部结构信息,通过最小化类内近邻分离度以及最大化类间近邻分离度来提高鉴别性能,从而进一步反映了数据的局部结构,提高了识别率。通过在ORL(Olivetti-Oracle Research Lab)和YALE两个标准人脸库上实验验证了该算法的有效性。Locality Preserving Projections（LPP） is an algorithm of local feature extraction,it can preserve local geometrical structure of the data set effectively.On the basis of LPP,Uncorrelated Discriminant Locality Preserving Projections（UDLPP） utilises label information,and achieves good discrimination capability by preserving the within-class geometric structure,while maximizing the between-class distance.A discriminant and analyse method of local structure preserving,called PCLSP,is proposed based on UDLPP and with the idea of UDLPP,which not only uses the label information of the data set but also takes the local information of the data set into account.It enhances discrimination capability by minimizing the scatter of within-class neighbourhood and maximizing the scatter of between-class neighbourhood,therefore further reflects the local structure information of the data set and increases the recognition rate.Experimental results on ORL and YALE face databases indicate that the proposed algorithm is effective.

关 键 词：LPP UDLPP流形 人脸识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]