白化主成分分析类算法在人脸识别中的应用

Whitening PCA-clas algorithm for face recognition

针对能量谱的不平衡性会影响人脸识别效果的问题,基于白化脸的概念提出了白化主成分分析类算法的框架。该算法框架使用1个白化滤波器和1个低通滤波器对原始图像进行预处理,然后结合传统的PCA类算法提取特征向量(或矩阵),最后通过k-NN分类方法进行人脸识别。利用ORL人脸图像库进行实验,实验结果表明该算法框架改善了人脸识别的效果,提高了识别的正确率。

The unbalanced power spectra of facial images will result in potential problems when used in face recognition.To solve this problem,a Whitening PCA-class algorithm framework based on the concept of whitenedfaces is proposed.In the algorithm framework,it preprocesses the original images by a whitening filter and a low-pass filter at first,then extract features vectors（or matrices） combined with the traditional PCA-class algorithm,and finally complete the face recognition through the k-NN classification method.The result on the experiments in the ORL face image database shows that the algorithm framework brings better recognition performance and achieves higher recognition accuracy.