基于LBP特征和贝叶斯模型的单样本人脸识别  

Single sample face recognition based on LBP feature and Bayes model

作　　者：杨军[1,2] 高志升[1] 袁红照[1] 余静[1] 张秀琼[1] 柳翠寅[1] 

机构地区：[1]四川大学计算机学院,四川成都610064 [2]四川师范大学计算机科学学院,四川成都610066

出　　处：《光电子．激光》2011年第5期763-765,共3页Journal of Optoelectronics·Laser

基　　金：国家"863"计划资助项目(2006AA12A104);国家自然科学基金资助项目(60736046)

摘　　要：针对单样本人脸识别这一人脸识别中的难点问题,提出了一种基于局部二元模式(LBP)直方图特征和贝叶斯(Bayes)模型的人脸识别方法。首先在独立的训练集上学习同类样本和异类样本的LBP直方图特征的相似度先验信息,估计同类和异类的类条件概率密度函数,在识别过程中利用一对图像的LBP直方图相似度计算该对图像属于同一类的后验概率,最后利用贝叶斯规则进行分类。在ORL和Yale数据库上的人脸识别实验表明,本文方法的人脸识别正确率相对LBP方法分别提高了3.1%和1.4%。It is difficult problem for face recognition on condition that only one image of a person included in face database.For dealing with the problem,we propose a face recognition method based on local binary pattern（LBP） histogram feature and bayes classifier.The method firstly learned the prior information of similarity of LBP histogram in the same class and in different class,and evaluated the same class conditional probability density function and different class conditional probability density function.When a probe image was discriminated,the method calculated the similarity of the probe and a template image in database using their LBP histogram features,and then evaluated the posterior probability of the pair images coming from the same person.Finally,the probe image was classified by bayes rule.The propose method fuses the LBP feature and prior information into face data.Experiment results on ORL and Yale databases show that it is valid and the recognition accuracy rates are improved by 3.1% on ORL database and by 1.4% on Yale database compared with LBP method.

关 键 词：人脸识别 局部二元模式(LBP) 贝叶斯(Bayes)模型 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]