基于新的Gabor变换与直接LDA方法的人脸识别  

Face Recognition Based on A New Gabor Transform Method and Direct LDA

作　　者：聂方彦[1] 张平凤[1] 

机构地区：[1]湖南文理学院,湖南常德415000

出　　处：《微计算机信息》2008年第15期277-279,共3页Control & Automation

摘　　要：提出一种基于新的Gabor滤波变换与直接线性判别分析的人脸识别方法(GDLDA)。首先应用一种新的Gabor小波对脸谱图像进行滤波变换,然后应用直接线性判别分析法(DLDA)提取经滤波变换后的脸谱Gabor变换特征进行识别。最后使用近邻算法在标准AT&T脸谱数据库上对GDLDA与DLDA方法进行脸谱分类识别比较,实验结果表明了GDLDA方法的有效性。A face recognition method was proposed. In this method, a new Gabor transformation was combined with direct linear discriminant analysis（DLDA）. Firstly, a new group of Gabor filter were convoluted with face images, then extracted the Gabor feature using DLDA, finally K-Nearest Neighbor algorithm was used for classification and recognition. Experimental results on AT＆T face database compared with DLDA illustrated the effectiveness of the proposed method.

关 键 词：人脸识别 GABOR滤波 直接线性判别分析 近邻算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]