基于特征脸和支持向量机的人脸识别技术研究  

Research on Face Recognition Technology Based on Eigen Face and Support Vector Machine

作　　者：王志杨[1] 刘金龙[2] 毛世伟[1] 

机构地区：[1]安徽电子信息职业技术学院信息工程系,安徽蚌埠233030 [2]淮海工学院电子工程学院,江苏连云港222005

出　　处：《淮海工学院学报（自然科学版）》2015年第1期23-29,共7页Journal of Huaihai Institute of Technology:Natural Sciences Edition

基　　金：安徽省高校省级优秀青年人才基金重点项目(2013SQRL104ZD)

摘　　要：在人脸检测的基础上,针对人脸识别问题,将特征脸理论和支持向量机相结合,把待识别样本投影到由特征脸张成的特征空间,从而提取出有效的可供识别特征,再将提取的特征送入支持向量机训练、分类,最终实现识别算法.该算法充分利用了主分量分析法在特征提取方面的有效性以及支持向量机在处理小样本问题和泛化能力强等方面的优势,使得算法具有较高的识别率.最后在设计的软件平台上进行了仿真,仿真结果表明,该系统可以运用在小型人脸识别的场合.On the basis of face detection,according to face recognition,this paper combines the theory of eigen face with support vector machine.It projects the samples to be recognized into feature space composed of eigen face,thereby extracting the effective feature for recognition.Then the feature is sent into support vector machine to train and classify.At last the recognition algorithm is achieved.The algorithm makes full use of the validity of principal component analysis in feature extraction and the advantage of support vector machine processing the problem of small samples and its strong generalization ability,which makes recognition rate of the algorithm higher.Finally the simulations are carried out on the designed software platform.The simulation result proves that the system can be applied in a miniature situation of face recognition.

关 键 词：人脸识别 特征脸 支持向量机 主分量分析 K-L变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]