计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (26): 183-185.DOI: 10.3778/j.issn.1002-8331.2008.26.056

• 图形、图像、模式识别 • 上一篇    下一篇

基于ICA和NFL与NN联合分类器的人脸识别

余慧海,申金媛,刘润杰   

  1. 郑州大学 信息工程学院,郑州 450052
  • 收稿日期:2007-11-07 修回日期:2008-01-28 出版日期:2008-09-11 发布日期:2008-09-11
  • 通讯作者: 余慧海

Face recognition with new classifier based on ICA features

YU Hui-hai,SHEN Jin-yuan,LIU Run-jie   

  1. College of Information and Engineering,Zhengzhou University,Zhengzhou 450052,China
  • Received:2007-11-07 Revised:2008-01-28 Online:2008-09-11 Published:2008-09-11
  • Contact: YU Hui-hai

摘要: 提出了一种基于最近邻特征线(NFL)与最近邻(NN)联合分类器进行人脸识别的方法。首先对人脸图像用主成分分析(PCA)降维,然后用快速独立变量分析(FastICA)提取独立基,分类时采用最近邻特征线和最近邻分类器的联合分类器进行分类。该方法综合了NFL和NN的优势,充分利用了同类之间相似,距离最短的性质。实验表明此方法提高了人脸识别率,是一种可行的人脸识别方法。

Abstract: A face recognition method based on Nearest Feature Line(NFL) combined with Nearest Neighbor(NN) is proposed.Firstly,the human faces data project to a low dimensional space with PCA,and then the features of faces are extracted by FastICA.At last faces are recognized by a classifier which is composed by NFL combined with NN.This method adequately uses the maximum similarity and shortest distance among the same person’s faces.The experiment results show that the method can get a high recognition rate.