改进的PCA算法在人脸识别中的应用研究

Face recognition with improved Principal Component Analysis

PCA将图像看作具有一定分布的随机向量,可以得到人脸图像分布的主成份方向,从本质上来讲这是一种基于统计分析信息的方法。由于人脸结构的相似形,本文提出一种改进的PCA算法,把输入图像经过直方图均衡化,边缘检测,突出人脸特征,得到降维的可对其进行PCA分析的图像。实验结果表明,改进的PCA算法在不降低识别率的前提下。

Principal Component Analysis recognized the image as a random vector with a certain distribution,It could produce prieipal component direction of the face images.In essence,it was based on statistical analysis of information. Because of the similar figures of the face structure,This paper presented an improved PCA algorithm.The face images were processed by histogram equalization ,edge detection in order to outstand face feature and get the lower-dimentional images,which is used for future normal PCA analyse . Experimental results show that the improved face recognition PCA algorithm greatly improve the speed, but the recognition rate do not decline.