作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2010, Vol. 36 ›› Issue (18): 188-190. doi: 10.3969/j.issn.1000-3428.2010.18.065

• 人工智能及识别技术 • 上一篇    下一篇

基于模糊积分分类器融合的人脸识别

顾晓敏,林锦国,梅 雪   

  1. (南京工业大学自动化与电气工程学院,南京 210009)
  • 出版日期:2010-09-20 发布日期:2010-09-30
  • 作者简介:顾晓敏(1985-),女,硕士研究生,主研方向:模式识别;林锦国,教授;梅 雪,博士
  • 基金资助:

    江苏省“六大人才高峰”基金资助项目(07-E-024)

Face Recognition Based on Classifiers Fusion Using Fuzzy Integral

GU Xiao-min, LIN Jin-guo, MEI Xue   

  1. (School of Automation & Electrical Engineering, Nanjing University of Technology, Nanjing 210009, China)
  • Online:2010-09-20 Published:2010-09-30

摘要:

提出一种基于模糊积分分类器融合的人脸识别算法。对人脸图像进行小波变换,选取合适的小波基函数及有效的分解层数,提取低频分量系数作为分类特征设计分类器。对原图像采用2DPCA进行特征提取设计另一分类器,采用模糊积分的算法融合2个分类器并得出最终分类识别结果。实验结果表明,模糊积分能够有效融合分类器的互补信息,提高系统的分类性能,从而提高人脸识别率。

关键词: 小波变换, 二维主成分分析, 模糊积分, 分类器融合, 人脸识别

Abstract:

This paper proposes a face recognition algorithm based on classifiers fusion using fuzzy integral. Wavelet transform is taken on after an appropriate mother wavelet and decomposition layer are chosen by theory and simulation, and the low-frequency component coefficients of the last layer are extracted to make one classifier. 2DPCA is applied to make the other. Fuzzy integral is applied to reach the final decision. Experimental results show the algorithm can obtain complementary information and improve the classification capbility so as to promote the recognition performance.

Key words: wavelet transform, 2DPCA, fuzzy integral, classifier fusion, face recognition

中图分类号: