基于复主分量分析的人脸识别

Face Recognition Based on Principal Component Analysis in the Complex Space

提出了一种基于复主分量分析的人脸识别新方法。首先采用两种不同的K L变换分别降低原始图像空间的维数,得到高维原始图像的两种简约表示。然后利用复向量将同一样本的两组特征向量合并在一起,通过运用复主分量分析,来抽取人脸图像的有效鉴别特征。最后在ORL人脸库上实验结果表明所提出的方法不仅识别性能优于经典的Eigenfaces和Fisherfaces方法,而且仅用27个特征识别率就达到96%。

Presents a novel face recognition method based on complex principal component analysis . First,in order to obtain two kinds of different simple expressions of the high orginal image vector,uses two kinds of KL transform methods to respectively reduce the dimensionalities of the original image vector space. Then,combines the two kinds of resulting features together by a complex vector,and employes the principal component analysis for feature extraction in the complex feature space. Finally,experimental results on ORL face database show that the proposed method not only is more effective than the classical Eigenfaces and Fisherfaces,but also achieves a recognition accuracy of 96% using only 27 features.