基于双向PCA和K近邻的人脸识别算法  

Face recognition algorithm based on BD-PCA and K-NN

作　　者：王心醉[1,2] 李岩[1] 郭立红[1] 肖永鹏[3] 董宁宁[1] 杨丽梅[4] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京100039 [3]东北师范大学计算机学院,吉林长春130117 [4]长春工业大学机电工程学院,吉林长春100012

出　　处：《解放军理工大学学报（自然科学版）》2010年第6期623-627,共5页Journal of PLA University of Science and Technology(Natural Science Edition)

基　　金：中国科学院知识创新工程领域前沿资助项目;长春市科技攻关项目(07163UC070)

摘　　要：针对当前人脸识别算法的运行速度和识别准确率的矛盾,提出一种基于双向主成分法(Bidirectional PCA,BD-PCA)和K近邻法K-NN(K-nearest neighbor)的人脸识别算法。在VC6.0平台下基于ORL人脸库进行实验,首先利用双向PCA算法对训练人脸样本和测试人脸样本进行方向和列方向降维并提取特征脸;然后用K近邻方法对特征脸进行人脸匹配。实验表明,提出的算法在有效降低运算时间的同时,又能取得很高的识别准确率,具有一定的可行性。Aiming at the contradiction between the running time and the accuracy of the existing face recognition algorithm,a face recognition algorithm based on BD-PCA and K-NN was proposed.The algorithm was tested on the ORL face database at VC 6.0.Firstly,the dimension of the row and the column of training and testing samples was reduced and the Eigenface was obtained using the BD-PCA algorithm.Then,the Eigenface was matched using the K-NN algorithm.The experimental result indicates that the algorithm can get high accuracy while greatly reducing the running time,which proves the feasibility of the algorithm.

关 键 词：双向PCA K-NN 人脸识别 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]