Journal of Computer Applications ›› 2005, Vol. 25 ›› Issue (05): 1102-1104.DOI: 10.3724/SP.J.1087.2005.1102

• Graphics, image and multimedia • Previous Articles     Next Articles

S-LISSOM based approach for face recognition

 YANG Zhi1, WANG Tao2   

  1. 1. School of Software, Tsinghua University, Beijing 100084, China; 2. Intel China Research Center, Beijing 100020, China
  • Online:2005-05-01 Published:2005-05-25

基于S-LISSOM模型的人脸识别方法

杨治1,王涛2   

  1. 1.清华大学软件学院; 2.英特尔集团中国研究中心
  • 基金资助:

    国家 863计划资助项目( 2003AA414031)

Abstract: A new approach to face recognition was proposed,based on Simplified Laterally Interconnected Synergetically Self-Organizing Map(S-LISSOM) Model and a EHMM. The approach extracted face character mapped from the gray image by using the S-LISSOM Model to form the Observing Vectors. Then the faces were classified by Hidden Markov Models. The experiment results were compared with other two HMM based methods. The results show that the method proposed has a lower face recognition error rate.

Key words: S-LISSOM, HMM, face recognition

摘要: 提出了一种利用视皮层认知模型识别人脸的方法。该方法建立一种简化的感受野切片连接自组织映射简化模型(S LISSOM),模拟人脸图像在大脑视皮层的映射特征,作为隐马尔可夫模型(HMM)的观测向量进行人脸识别。实验结果同其他特征的人脸识别方法进行了比较,该方法更加有效的提取的人脸特征,提高了人脸识别率。

关键词: S-LISSOM, 隐马尔可夫模型, 人脸识别

CLC Number: