摘要
针对非线性特征提取问题,基于核最大间距准则(KMMC),提出一种新的特征提取方法,即一组具有统计不相关性的最优核鉴别矢量集的简单计算方法.与原KMMC特征提取方法相比,新的特征提取方法消除了最优核鉴别矢量间的统计相关性,提高了特征提取的有效性.通过在ORL人脸库和YALE人脸库上进行试验,结果表明提出的特征提取方法在有效性方面整体上好于原KMMC特征提取方法和常用的核主成分分析(KPCA)法.
A new feature extraction method based on kernel maximum margin criterion （KMMC） Was presented for nonlinear feature extraction which is a simple algorithm of statistically uncorrelated optimal discriminant vectors in kernel feature space. Compared with the original KMMC feature extraction method, the proposed method is powerful in eliminating the statistical correlation between feature vectors and improving efficiency of feature extraction in the high dimensional feature space. The experimental resuits on Olivetti Research Laboratory（ORL） face database and YALE face database show that the new method is better than original KMMC and kernel principal component analysis （KPCA） in terms of efficiency and stability about feature extraction.
出处
《江苏大学学报（自然科学版）》
EI
CAS
北大核心
2008年第5期441-444,共4页
Journal of Jiangsu University：Natural Science Edition
基金
江苏省高校自然基金资助项目(06KJD520085)
兰州商学院高层次人才基金资助项目(4086)
南京林业大学人才基金资助项目(2002-10)
关键词
核最大间距准则
最优核鉴别矢量
特征提取
统计不相关性
人脸识别
kernel maximum margin criterion
optimal kernel discriminant vectors
featui-e extraction
statistically uncorrelation
face recognition
