基于频谱的人脸识别方法  

Face Recognition Based on Frequency Spectrum

作　　者：苏宏涛[1] 赵荣椿[1] 

机构地区：[1]西北工业大学计算机科学与工程系,陕西西安710072

出　　处：《计算机应用研究》2003年第7期67-68,74,共3页Application Research of Computers

摘　　要：小波变换后的低频子带图像既去除了某些表情变化,又减小了数据量,而图像的频谱特征则具有良好分类特性,因此两者结合后得到的频谱脸在人脸识别方面具有相当高的应用价值。先利用小波变换和Fourier变换求得原始人脸图像的频谱脸(Spectrofaces),再对频谱脸继续求取各自的本征脸(Eigenface)和LDA(LinearDiscriminantAnalysis)特征作为分类特征,并利用了不同的分类方法进行识别。实验是利用ORL人脸库进行的,实验结果证明了比起直接利用空间域上原始图像的识别方法来说,基于频谱的方法可以有效提高识别率。The wavelet transformation not only diminishes the expression variation also reduces the image size in the low frequency subband image,and the classification quality of the image frequency spectrum feature is very good,so the spectrum face acquired by combining the wavelet and frequency spectrum has great merit in face recognition. In this paper,the spectrum face of original image was computed via wavelet transform and Fourier transform,then the eigenface and LDA features of spectrum face were extracted as classification feature. At last,several classification methods were adopted to get the recognition result. The experiment was carried out using ORL face database,and the experiment result proved that the method based on frequency spectrum could effectively improve the recognition rate comparing to the method using original image in spatial field.

关 键 词：人脸识别 本征脸 LDA 频谱脸 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]