基于Gabor-RSC的人脸识别算法  

FACE RECOGNITION ALGORITHM BASED ON GABOR-RSC

作　　者：陈志涛[1] 李军[2] 杨小东[1] 

机构地区：[1]顺德职业技术学院,广东佛山528000 [2]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《计算机应用与软件》2015年第9期181-183,218,共4页Computer Applications and Software

基　　金：顺德科技计划项目(20120202088)

摘　　要：为了提高人脸识别的准确率,提出一种基于Gabor特征和鲁棒稀疏表征相融合的人脸识别算法(Gabor-RSC)。首先采用Gabor滤波器提取人脸图像的多尺度和多方向特征,并采用主成分分析降低特征维数,然后采用鲁棒稀疏编码算法对人脸进行识别,最后采用Yale和ORL人脸库进行仿真测试。结果表明,Gabor-RSC算法提高了人脸的识别正确率,鲁棒性更高。In order to improve face recognition accuracy, the paper proposes a face recognition algorithm, it is based on the combination of Gabor feature and robust sparse representation （Gabor-RSC）. First, it uses Gabor filter to extract multi-scale and multiple orientation features of face images, and employs principal component analysis to reduce the features dimensionality; then it uses robust sparse coding algorithm to recognise the face; finally, it conducts simulation tests on Yale and ORL face databases. Results show that the proposed Garbor-RSC algorithm improves face recognition accuracy, and has higher robustness.

关 键 词：特征提取 迭代加权 稀疏表示 人脸识别 

分 类 号：TP317[自动化与计算机技术—计算机软件与理论]