Abstract: This paper proposes a face recognition algorithm based on both Bit planes and generalized PCA. With Bit planes of face images, it gets features. Generalized PCA is applied to the new face images constructed by feature fusion. It is shown through experiments that, the proposed approach can not only increase the recognition rate, but also need fewer features to reach its stable recognition rate.

Key words: Face recognition, Bit planes, Feature fusion, Generalized PCA

摘要： 提出了一种结合Bit平面信息和广义PCA进行人脸识别的新算法。利用人脸图像的Bit平面信息，经特征融合来构造新的人脸，在此基础上再进行广义PCA分析。实验表明，该文提出的方法不仅能提高人脸的识别率，而且在人脸特征空间的维数较低时，识别率已经达到稳定。

关键词: 人脸识别, Bit平面, 特征融合, 广义PCA