In face recognition algorithms,the increase of feather dimensionality has always over⁃burdened the algorithm operation, so a new face recognition algorithm based on UMPCA and LDA is proposed.While the algorithm reduces the dimensionality,it remains the inner structure information as much as possible.UMPCA seeks a tensor⁃to⁃vector projection that captures most of the variation in the original tensorial input while obtaining uncorrelated features through successive variance maximization.A subset of features extracted is processed by classical LDA to find the best subspaces.Finally,the comprehensive experiments are provided on AT&T databases and the experiment results show its performance over other PCA plus LDA based algorithms.%针对在人脸识别算法中,维数的增加往往会给算法的运算带来沉重负担的问题,提出了一种新的基于非相关多线性主成分分析( UMPCA)和线性判别分析( LDA)的人脸识别算法,算法在保证在降维的时候保留尽可能多的内部结构信息。 UMPCA通过一张量至向量的过程,可直接获取原张量数据的绝大部分非相关特征,提取的特征再通过经典算法LDA处理。利用AT&T人脸数据库对该算法进行了实验,实验数据分析显示该算法优于其他同类算法。
展开▼