结合小波低频子带的主成分分析方法  

Principal Component Analysis Combined with Wavelet Low-Frequency Band

作　　者：何家忠[1] 杜明辉[1] 

机构地区：[1]华南理工大学电子与信息学院,广东广州510640

出　　处：《华南理工大学学报（自然科学版）》2007年第1期44-48,共5页Journal of South China University of Technology(Natural Science Edition)

基　　金：广东省自然科学基金资助项目(05006593)

摘　　要：针对单样本人脸识别问题,提出了一种结合小波低频子带的主成分分析方法.为了加强单样本图像的分类信息,该方法将原训练样本与其小波低频子带重构图相结合,然后对结合的训练样本进行主成分分析.在ORL人脸库上的实验结果表明,当训练集中每个人只有一幅人脸样本图像时,文中方法的识别率比标准特征脸法高3.6%,而所使用的特征脸个数减少14.8%.This paper proposes a principal component analysis method combined with wavelet low-frequency band to deal with the face recognition with single training sample. To enhance the classification information of a single sample image, the method combines the original training image with its reconstructed image based on wavelet low-frequency band, and performs the principal component analysis on the joined version of the training image. Experimental results on the ORL database show that, when each person has only one training sample, the proposed method achieves 3.6% higher recognition accuracy and uses 14. 8% fewer eigenfaces than the standard eigenface algorithm.

关 键 词：人脸识别 小波变换 主成分分析 特征脸 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]