摘 要：人脸自动识别是计算机模式识别领域的一个活跃课题，有着十分广泛的应用前景。为了对人脸图像的特征向量进行分类已达到人脸识别的目的，本文提出了运用BP神经网络进行人脸识别的方法。 

　　　将BP神经网络用于人脸识别，并建立了人脸识别模型，该识别模型包括图像压缩、图像抽样、输入矢量标准化、BP神经网络与竞争选择处理过程、具有简单，识别率高的特点。BP神经网络具有正向传播和反向传播的特性，从而保证了分类的准确性，所以本文主要研究如何在MATLAB中把BP神经网络应用在人脸的识别分类上。

关键词：图像压缩  图像抽样  神经网络  人脸识别

Abstract：The automatic facial recognition is one of hot topics in computer identification field and has a wide range of applications.In order to classify the eigenvectors of human face with facial recognition met，the paper raises the method of using BP neural networks to do facial recognition.

　　The paper applies BP neural networks in facial recognition and builds the model of facial recognition. The recognition model included image compression，image sampling，input vector standardization. The BP neural networks and competitive selection process are simple and of high identifying ratios.

Key Words：image compression  image sampling  neural networks  facial recognition