| 二维PCA人脸特征提取算法及其改进 |
| |
| 引用本文： | 沈银银,冯颖凌. 二维PCA人脸特征提取算法及其改进[J]. 南通职业大学学报, 2010, 24(3): 98-100. DOI: 10.3969/j.issn.1008-5327.2010.03.027 |
| |
| 作者姓名： | 沈银银 冯颖凌 |
| |
| 作者单位： | 1. 南通大学理学院,江苏南通226007;南通科达市政交通设计有限公司,江苏南通226001
2. 南通大学理学院,江苏南通,226007 |
| |
| 摘 要： | 基于DiagPCA（对角主成分分析）及平均脸的方法对二维主成分分析（2DPCA）方法进行了改进,既考虑到构造2DPCA训练样本人脸间散布矩阵时使特征最大化,减少了同类人脸之间的特征差异,又利用图像矩阵对角化将图像的行、列关系联系起来,并利用ORL人脸数据库进行实验。结果显示,该方法可提高人脸识别率,且降低了特征提取的时间。
|
| 关 键 词： | 二维主成分分析 特征提取 人脸识别 DiagPCA 对角平均脸 |
Two-dimensional PCA Feature Extraction Algorithm and Its Improvement in Face Recognition |
| |
| SHEN Yin-yin,FENG Ying-ling. Two-dimensional PCA Feature Extraction Algorithm and Its Improvement in Face Recognition[J]. Journal of Nantong Vocational College, 2010, 24(3): 98-100. DOI: 10.3969/j.issn.1008-5327.2010.03.027 |
| |
| Authors: | SHEN Yin-yin FENG Ying-ling |
| |
| Affiliation: | 1.College of science,Nantong university,Nantong 226007,China;2.Nantong Keda municipal transportation design Co.,LTD,Nantong 226001,China) |
| |
| Abstract: | Two-dimensional principal component analysis（2DPCA） method is one of the methods which are often used in face recognition.This paper has improved 2DPCA basing on the methods of average faces and diagonal 2DPCA,and this method does not only consider making features max when we tectonic the scatter matrix of training sample faces of 2DPCA,but also reduces the feature differences between the same-class faces,and connects rows and column of the image matrix by making the image matrix to diagonalization.And we did experiments on the ORL face database,it shows that this method has really improved the recognition. |
| |
| Keywords: | DiagPCA |
| 本文献已被 维普 万方数据 等数据库收录！ |
|
