基于先验概率与不同小波基的人脸识别研究

The human face recognition base on prior probability and diffrent wavelet bases

由于小波分解在空域和频域上都能提供良好的局部信息,尤其是在小波分解后可以减少图像的分辨率,进而相应地减少计算复杂度,因此小波变换经常用于图像处理和图像分析中。通过实验得出了基于各种小波基的人脸识别的识别率和识别速度,并根据各种小波基的识别率和识别速度这一先验概率,将各种小波基在识别时进行融合,提出了一种新的人脸识别方法。

Wavelet transform（WT） has been a very popular tool for image analysis,because WT have the advantages of good time and frequency localizations, and the resolutions of the subband images are reduced by decomposing an image using WT. In turn, the computational complexity will be reduced dramatically by working on a lower resolution image. There are the different recognition ratios and recognition velocity base on the different wavelet bases. And find a new algorithm of human face recognition.