摘要
研究并实现了利用Gabor滤波器和Fisher线性鉴别分析(FLDA)方法的动态人脸识别考勤系统。系统实现的基本思想是运用Gabor变换提取人脸的局部特征和经过Gabor处理后使得人脸对光照变化不敏感;进一步利用FLDA来降维和隐含地提取最有利于分类的最佳鉴别特征;最后将视频采集的考勤图像与训练库中的图像通过比对,得出识别结果。实验结果表明,利用该方法开发的动态人脸识别考勤系统具有识别率高、实用性好、可靠性强等特点。
Dynamic Face Recognition based Attendance System（DFRAS） by using Gabor filter and Fisher Linear Discriminant Analysis（FLDA） is realized successfully. In this paper, the algorithm＇s idea for DFRAS is that human local feature is extracted by using Gabor conversion, furthermore, the best discriminant features conducive to classification are extracted by FLDA, and then recognition outcomes are come from comparing video face image with trained image base. Experimental results indicate that this method for Face Recognition based Attendance System has the advantages of high recognition rate, good practicability, robustness and so on.
出处
《计算机科学》
CSCD
北大核心
2008年第4期294-296,F0003,共4页
Computer Science
基金
国家自然科学基金(60272095)
教育部博士点基金(20020610013)
关键词
考勤系统
视频序列
人脸识别
GABOR滤波器
FLDA
特征提取文献标识码
Attendance system, Video sequence, Face recognition, Gabor filter, Fisher linear discriminant analysis, Feature extraction
