摘要： 提出基于二维Gabor小波特征(Gaborface)矩阵的人脸表征方法，二维Gaborface矩阵有别于常规的一维采样特征矢量方法。对该方法的2种运用方式：整体Gaborface表征(EGFR)和多通道Gaborface表征(MGFR)进行了研究，在ORL数据库中采用二维PCA人脸识别算法进行了实验。对比实验结果表明了该方法的有效性和可行性，特别是基于MGFR的2DPCA方法实现了100%的识别率。
关键词:
人脸识别,
Gabor小波特征,
2DPCA算法,
特征提取,
多通道,
决策级融合
Abstract: This paper introduces a novel face representation method based on two-dimensional Gabor wavelet feature (Gaborface) matrices instead of the conventional transformed one dimensional feature vectors. Both the ensemble Gaborface representation and the multi-channel Gaborface representation are studied and experimented on the ORL face database using two-dimensional principal component analysis face recognition approach (2DPCA). Experimental results show the efficiency and feasibility of the method, particularly, the 100% recognition rate is achieved by the MGFR-based 2DPCA.
Key words:
face recognition,
Gaborface,
Two-dimension Principal Component Analysis(2DPCA),
feature extraction,
multi-channel,
decision- level fusion
中图分类号:
王 琳;李勇平;王成波;张鸿洲. 基于二维Gabor小波矩阵表征人脸的识别算法[J]. 计算机工程, 2008, 34(2): 8-10.
WANG Lin; LI Yong-ping; WANG Cheng-bo; ZHANG Hong-zhou. Face Recognition Algorithm Based on Two-dimensional Gabor Wavelet Feature Matrices[J]. Computer Engineering, 2008, 34(2): 8-10.