一种用于取证的加权支持向量机的人脸识别方法  

A Face Recognition Method Based on Weighted Support Vector Machine Used for Forensics

作　　者：周建华[1] 徐联华[1] 史伟奇[1] 

机构地区：[1]湖南警察学院,湖南长沙410138

出　　处：《计算机时代》2010年第6期18-20,共3页Computer Era

基　　金：湖南省教育厅项目(08C027);2007湖南省公安厅科研计划项目批准立项课题;公安部科技创新项目(2005YYCXHNST095);湖南省高校优秀青年科学研究项目(07B017)

摘　　要：人脸识别是指计算机将检测到的人脸图像,与计算机中已存储的人脸进行匹配以认定身份的过程。根据公安取证的需要,并针对已有的支持向量机人脸识别技术的不足,提出了用于取证的加权支持向量机人脸识别方法。对ORL人脸数据集的检测试验表明,该方法不但可行,而且准确性及效率较高。Face recognition refers to the process of computer matching a detected face image against the faces stored in computer for identification. According to the need of police collecting evidences, and aiming at the disadvantages of existing face recognition technology with support vector machine, the face recognition method with weighted support vector machine for forensics is presented. The detection experiments on ORL face datasets show that this method is not only feasible, but also of higher accuracy and efficiency.

关 键 词：加权支持向量机 人脸识别 取证 加权映射 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]