计算机工程与应用 ›› 2007, Vol. 43 ›› Issue (36): 226-229.

• 工程与应用 • 上一篇    下一篇

基于BP神经网络的主分量分析人脸识别算法

赵立强1,张晓华1,2,高振波3,张洪亮1   

  1. 1.河北科技师范学院 数理系,河北 秦皇岛 066004
    2.哈尔滨理工大学 计算机科学与技术学院,哈尔滨 150080
    3.河北科技师范学院 计算机系,河北 秦皇岛 066004
  • 收稿日期:1900-01-01 修回日期:1900-01-01 出版日期:2007-12-21 发布日期:2007-12-21
  • 通讯作者: 赵立强

Face recognition based on BP nural networks and principle component analysis

ZHAO Li-qiang1,ZHANG Xiao-hua1,2,GAO Zhen-bo3,ZHANG Hong-liang1   

  1. 1.Department of Mathematics and Physics,Hebei Normal University of Science and Technology,Qinhuangdao,Hebei 066004,China
    2.College of Computer Science and Technology,Harbin University of Science and Technology,Harbin 150080,China
    3.Department of Computer,Hebei Normal University of Science and Technology,Qinhuangdao,Hebei 066004,China

  • Received:1900-01-01 Revised:1900-01-01 Online:2007-12-21 Published:2007-12-21
  • Contact: ZHAO Li-qiang

摘要: 提出了基于BP神经网络的主分量人脸识别算法。该算法首先用小波变换对人脸图像进行小波分解,形成低频小波子图,然后用主分量分析法构造特征脸子空间,将人脸图像在特征空间的投影作为BP神经网络的输入,由BP神经网络和后验概率转换器构成人脸识别器。针对ORL人脸库的实验结果表明该方法具有较高的识别率。

关键词: 人脸识别, BP神经网络, 主分量分析(PCA), 小波变换

Abstract: BP neural network combined with principal component analysis is applied to human face recognition.After extracting low frequency sub-band of face image in wavelet transform,the eigenface space is constructed by PCA.Then all samples are projected into the subspace,the coefficient of every sample is inputted to BP neural network,and the face recognizer consists of BP neural network and post-probability converter.The experiments on ORL face database indicate the recognition ratio is greatly improved.

Key words: face recognition, BP neural network, principle component analysis(PCA), wavelet transform