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.
出处
《上海大学学报（自然科学版）》
CAS
CSCD
北大核心
2006年第3期244-248,255,共6页
Journal of Shanghai University:Natural Science Edition
关键词
考勤系统
人脸识别
正、侧面图像
特征矢量
attendance system
face recognition
facial frontal and profile image
feature vector
