基于贝叶斯分类在人脸识别应用中的研究与实现  

Study And Realization of Bayes Classification Used in Face Recognition

作　　者：连国云[1] 刘辉[1] 邓彬[1] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650051

出　　处：《山西电子技术》2007年第6期38-39,42,共3页Shanxi Electronic Technology

摘　　要：本征脸法将图像看做矩阵,计算本征值和对应的本征向量作为代数特征进行识别。将本征脸法分别与欧氏距离判据及贝叶斯判据相结合进行人脸识别,并对两种判据进行了比较。此外还对贝叶斯判据在人脸识别应用中进行了改进,采用ORL图像库的实验表明,改进是有效的。Taking image as a matrix, an eigenface algorithm uses eigenvalues and corresponding eigenvectors in recognition. In this paper, it rises eigenface algorithm to make face recognition differently combined with Euclidian rule and hayes rule, and has cartied on the comparison with that two kinds of rules. In addition, it has made the improvement to bayes rule in the face recognition application. The face recognition experiment with the ORL face databases shows that the improvement is effective.

关 键 词：人脸识别 本征脸法 欧氏距离分类 贝叶斯分类器 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]