A fusion of WPCA and WLDA for face recognition based on dynamic feedback (DFWPCA+WLDA) is developed.This method firstly implements PCA dimension reduction to obtain the projection matrix,and then utilizes continuous feedback information to obtain the weighted value and weight the covariance matrix to optimize the projection matrix. Finally it adopts weighted LDA further to extract classification features.This dynamic feedback can make good use of useful information,and the weighted LDA can implement better classification.The experiment results on ORL and YALE face database show that this method is effective and has better perform-ance compared with PCA+LDA and WPCA+WLDA.%提出了一种基于动态反馈的融合加权主成分分析（ WPCA）和加权线性判别分析（ WLDA）的人脸识别方法（DFWPCA＋WLDA）。该方法首先进行主成分分析（PCA）降维得到投影矩阵，然后通过不断的反馈信息得到权值，从而加权协方差矩阵，优化投影矩阵，最后采用加权线性鉴别分析（ LDA）进一步提取分类特征。动态反馈能很好地利用样本的有用信息，加权LDA还能做到更好的分类。在ORL和YALE人脸库上的实验表明，该方法有效且性能优于PCA＋LDA和WPCA＋WLDA。
展开▼
