(完整版)基于matlab的人脸识别系统设计毕业设计

毕 业 设 计 [论 文]

题 目： 基于MATLAB的人脸识别系统设计

学 院： 电气与信息工程学院

专 业： 自动化

姓 名： 张迎

指导老师： 曹延生

完成时间： 2013.05.28

摘要

人脸识别是模式识别和图像处理等学科的一个研究热点，它广泛应用在身份验证、刑侦破案、视频监视、机器人智能化和医学等领域，具有广阔的应用价值和商用价值。人脸特征作为一种生物特征，与其他生物特征相比，具有有好、直接、方便等特点，因此使用人脸特征进行身份识别更易于被用户所接受。

人脸识别技术在过去的几十年得到了很大的发展，但由于人脸的非刚性、表情多变等因素，使得人脸识别技术在实际应用中面临着巨大的困难。本文针对近年来国内外相关学术论文及研究报告进行学习和分析的基础上，利用图像处理的matlab实现人脸识别方法，这种实现简单且识别准确率高，但其缺点是计算量大，当要识别较多人员时，该方法难以胜任。

利用MATLAB实现了一个集多种预处理方法于一体的通用的人脸图像预处理仿真系统，将该系统作为图像预处理模块可嵌入在人脸识别系统中，并利用灰度图像的直方图比对来实现人脸图像的识别判定。

关键词： 图像处理, Matlab, 人脸识别, 模式识别

ABSTRACT

Human face recognition focuses on pattern recognition ,image processing andother subjects.It is widely used in authentication,investigation,video surveillance,intelligent robots,medicine and other areas.Facerecognition has wide application and business value.Facial feature asabiological characteristic,compared with others is direct,friendly andconvenient.Facial featureemployed in authentication are user-friendly.

The technology of face recognition in the past few years obtained the very big development, but due to the face of nonrigid, expression and changeablefactors, the face recognition technology in practical application are facing great difficulties. This paper aimed at home and abroad in recent years the relevant papers and researchreports on study and on the basis of the analysis, some units within the data sensitivity places need to enter personnel to carry out limitation design and develop a set of identity verification identification system, the system uses PCA face recognition method, therealization is simple and the accuracy rate of recognition is high,but itsdrawback is that a large amount of calculation, when to identify more staff,this method is difficult to do.

The realization of a set of various pretreatment methods in one of the generic face image preprocessing simulation system based on MATLAB, the system is used as the image preprocessing module can be embedded in a face recognition system, and using the histogram matching gray image to realize the recognition of human face images to determine.