基于框架变换和PCA分析的人脸识别方法及实现  

Face Recognition Method Based on Tight Framelet Transformation and PCA

作　　者：王海辉[1] 施周虎[1] 邓俊军[2] 彭立中[2] 

机构地区：[1]北京航空航天大学数学系,数学,信息与行为教育部重点实验室,北京100083 [2]北京大学数学科学学院,北京100871

出　　处：《工程数学学报》2008年第4期571-575,共5页Chinese Journal of Engineering Mathematics

基　　金：国家自然科学基金重点项目(10426004;10701008)

摘　　要：研究了利用紧框架变换和主成分分析获取人脸特征,设计分类器进行分类识别,通过边缘信息和弱分类器的不同权值对分类识别进行优化。基于人脸数据库的实验结果表明,我们的方法是一种快速、有效和稳定的人脸识别方法。Framelet transformation and principle component analysis were employed to obtain facial features, and receptively identification was achieved through classifier designation. The improvement was made through the selection of different weights between the edge information and the weak classifier. Experimental results of face-based database show that our approach is a fast, effective and stable method for face recognition.

关 键 词：紧框架 主成分分析 人脸识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]