通过对主成分分析法(PCA)的数学公式进行改进,使其具有灰度归一化操作能力,从而克服光照对目标的影响,再将改进后的主成分分析法和Fisher线性判别分析方法组合起来用于人脸识别,在ORL人脸数据库上进行了实验,取得了满意的识别效果.
Presented an improved algorithm of PCA. It can solve the affection problem of image gray using withdraw. This algorithm was combined with Fisher Linear Distinguish Analysis. Then the combination has been used in face recognition. The experiment results give the effect support to the algorithm from ORL.