人脸识别考勤系统中有效人脸特征提取  

Efficient Feature Extraction in an Attendance System Based on Face Recognition

作　　者：张婷[1] 朱秋煜[1] 朱军林[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《上海大学学报（自然科学版）》2006年第3期244-248,255,共6页Journal of Shanghai University:Natural Science Edition

摘　　要：提出了一种适用于考勤系统的人脸识别算法.该算法首先从2台摄像机实时采集的正、侧面图像中分割出人脸区域,然后对图像中的人脸进行有效几何特征的提取,并构造出特征矢量与持卡者所持有的卡片中信息进行比对.经实验验证本系统算法简单、定位准确、识别率较高.A face recognition method applicable to attendance systems is proposed. Areas of a face image are first segmented from facial frontal and facial profile images that are dynamically captured with a camera. Effective geometric feature points extracted from the face regions are then used to construct feature vectors. These feature vectors are compared with the vector information in the card held by the subject. The algorithm is simple, and a high recognition rate has been achieved in the experiments.

关 键 词：考勤系统 人脸识别 正、侧面图像 特征矢量 

分 类 号：TN911.73[电子电信—通信与信息系统] TP391.41[电子电信—信息与通信工程]