基于环形对称Gabor和双边2DLDA的人脸识别研究

本文提出一种基于环形对称Gabor变换和双边线性判别2DLDA的人脸特征提取算法。首先采用环形对称Gabor变换进行人脸图像的特征提取,用两种不同的融合方案得到特征图像,再利用双边线性判别2DLDA算法对特征图进行数据降维,最后使用最近邻分类法分类。在ORL和PIE人脸库中的实验结果表明,本文提出的算法可以有效降低冗余度、减少耗时,提高人脸识别率。

This paper proposes a face feature extraction algorithm based on circular symmetric Gabor transform and bilateral linear discriminant 2DLDA. Firstly, the feature extraction of face images is carried out by using circular symmetric Gabor transform. The feature images are obtained by two different fusion schemes. Then the dimension of data is reduced by the bilateral linear discriminant 2DLDA algorithm. Finally, the nearest neighbor classification method is used to accomplish classification. The experimental results in ORL and PIE face database show that the proposed algorithm can effectively reduce redundancy, reduce time consumption and improve face recognition rate.