基于小波分析和几何特征的人脸识别方法研究

A research in face recognition based on wavelet analysis and geometric features

在主成分分析法(PCA)和独立成分分析法(ICA)等理论基础上,提出一种结合人脸几何特征和独立Gabor小波特征分析的人脸识别方法。在对人脸图像进行二维小波分解的基础上,从人脸图像的下采样Gabor小波图像中得到一个Gabor小波特征向量并利用PCA法降维,在ICA法的基础上得到独立Gabor小波特征,并结合人脸面部器官的位置和轮廓及器官距离等所构成的几何特征进行人脸识别。

Based on the theorem of principal component analysis and independent component analysis, an algorithm of human face recognization combined independent Gabor features with face geometric features is proposed. Firstly, deeompound the face images with wavelet of two dimensions ,Secondly, acquire a Gabor feature vector based on Gabor wavelet downsamples of face images and reduce the veetor＇s dimensions. Based on independent component analysis, acquire independent Gabor features. Finally, obtain the face geometric features such as the position and contour of face apparatus and the distances among them, then make full use of the information obtained from the face geometric features and independent Gabor features to recognize the faces.