基于小波变换与2DPCA的人脸特征提取方法

提出一种基于小波变换与2DPCA(二维主元分析)的人脸特征提取方法.首先用不同滤波嚣对人脸图像进行小波变换,选取合适的小波基对训练图像进行两次小波分解,然后采用以K-L变换为基础的二维主元分析方法来进一步获取人脸特征信息,最后采用最近邻分类法对特征进行分类.该方法计算简单,运算速度快,比较适合实时系统的要求.实验结果表明,该方法是一种比较有效的人脸特征提取方法.

Face Feature Extraction Based on Wavelet Transform and 2DPCA

A method of face feature extraction is proposed.Firstly,the face images are wavelet transformed with different filters.Then we transform the images with appropriate wavelet - base to get the coefficients of the sub - images. Coefficient matrixes of different sub-images are used as the input of the 2DPCA based on K-L change.Finally,the derived eigenvectors are used to classify the images.The nearest neighbor classifier is employed for face classification; This method can calculate easily and rapidly.The simu...