FACE RECOGNITION RESEARCH BASED ON ANTI-SYMMETRICAL WAVELET AND EIGENFACE

摘要：

In this paper, a new human face recognition method based on anti-symmetrical biorthogonal wavelet transformation (ASBWT) and eigenface was proposed.First the anti-symmetrical biorthogonal wavelet is chosen to degrade the face image dimension, meanwhile complete the process of face location and segmentation; And then human face is reverted through the face space of Eigenface, the traditional average human face is replaced in the within-class scatter matrix.This within-class scatter matrix is used to calculate within-class and between-class distance proportion as a rule function, calculate the twice eigenface through Discrete Karhunen-Loeve Transform (DKLT), and use Singular Value Decomposition (SVD) method to calculate the eigenvector.Finally we compute the weights and classify the face images.The results show that the proposed method has higher recognition rate and more robust than the traditional eigenfaee analysis method.

关键词： Face recognition Anti-symmetrical biorthogonal wavelet Eigenfaee SVD

作者: QIAN YIN ZHI-YONG YUAN YING KONG PING GUO

作者单位: Image Processing and Pattern Recognition Laboratory, Beijing Normal University, Beijing 100875, Chin School of Computer Science, Wuhan University, Wuhan 430079, China Image Processing and Pattern Recognition Laboratory, Beijing Normal University, Beijing 100875, Chin

会议类型: 国际会议

会议名称: 2007 International Conference on Machine Learning and Cybernetics(IEEE第六届机器学习与控制论国际会议)

会议地点: 香港

会议语种:英文

页码: 366-371

在线出版日期: 2007-08-19（万方平台首次上网日期，不代表论文的发表时间）