首页 | 本学科首页   官方微博 | 高级检索  
     

基于低秩特征脸与协同表示的人脸识别算法
引用本文:杨明中,杨平先.基于低秩特征脸与协同表示的人脸识别算法[J].液晶与显示,2017,32(8):650-655.
作者姓名:杨明中  杨平先
作者单位:四川理工学院 自动化与信息工程学院, 四川 自贡 643000
摘    要:在人脸识别中,人脸图像往往受到表情、光照、遮挡、姿态变化的影响,对此本文提出一种基于低秩特征脸与协同表示的人脸识别算法。该算法先用低秩矩阵恢复算法分解出训练样本图像的误差图像,再分别对训练样本与误差图像提取特征构造特征字典,计算测试样本图像特征字典下的协同表示系数,最后通过重构误差进行分类。通过AR和ORL人脸库进行实验,结果表明,本文提出的人脸识别算法的识别率、识别速率得到有效提高。

关 键 词:人脸识别  低秩矩阵  特征脸  协同表示
收稿时间:2017-01-17

Face recognition algorithm based on low-rank eigenface and collaborative representation
YANG Ming-zhong,YANG Ping-xian.Face recognition algorithm based on low-rank eigenface and collaborative representation[J].Chinese Journal of Liquid Crystals and Displays,2017,32(8):650-655.
Authors:YANG Ming-zhong  YANG Ping-xian
Affiliation:School of Automation & Information Engineering, Sichuan University of Science & Engineering, Zigong 643000, China
Abstract:For face recognition, the face image tend to get variations of expression, lighting, occlusion and pose. In the paper, we propose a face recognition algorithm based on low-rank eigenface and collaborative representation. In the algorithm, we firstly obtain the error images of the training images using the low-rank matrix recovery algorithm. Then, we calculate the feature vectors of the training images and the corresponding error images to constitute a feature dictionary respectively, and calculate collaborative representation of test image with feature dictionary. Finally, the test image was classified by the reconstruction error. Experimental results on AR and ORL databases show that the proposed face image recognition algorithm has a higher recognition rate and fast.
Keywords:face recognition  low-rank matrix  eigenface  collaborative representation
本文献已被 CNKI 等数据库收录!
点击此处可从《液晶与显示》浏览原始摘要信息
点击此处可从《液晶与显示》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号