基于判别邻域嵌入的人脸识别
Discriminant neighborhood embedding for face recognition
王国强 1欧宗瑛 1王海燕 2苏铁明3
作者信息
- 1. 大连理工大学精密与特种加工教育部重点实验室,辽宁,大连,116024;大连理工大学机械工程学院CAD&网络研究所,辽宁,大连,116024
- 2. 焦作市龙光影视设备有限责任公司,河南,焦作,454000
- 3. 大连理工大学机械工程学院CAD&网络研究所,辽宁,大连,116024
- 折叠
摘要
提出了一种人脸识别子空间方法:判别邻域嵌入(DNE).在框架中,训练样本数据的邻域和类关系被用来构建低维嵌入流形.在嵌入低维子空间后,同类样本保持它们固有的邻域关系,相反不同类近邻样本彼此远离.在ORL和Yale人脸数据库上,对提出的方法和主成分分析(PCA)、线性判别分析(LDA)、保持邻域嵌入(NPE)和保持局部投影(LPP)方法进行了比较,结果表明,提出的方法是有效的.
Abstract
A novel subspace method,called discriminant neighborhood embedding(DNE) ,is proposed for face recognition. In our framework,the neighbor and class relations of training samples data are used to construct the low-dimensional embedding submanifold. After being embedded into a low-dimensional subspace, the samples of same class maintain their intrinsic neighbor relations,whereas the neighboring samples of different class are far from each other. The proposed method was compared with PCA、LDA、NPE and LPP methods on the ORL and Yale face databases. Experimental results indicate the proposed method is effective.
关键词
子空间分析/判别邻域嵌入(DNE)/流形学习/降维/人脸识别Key words
subspace analysis/discriminant neighborhood embedding(DNE)/manifold learning/dimensionality reduction/face recognition引用本文复制引用
出版年
2008
NETL
NSTL
维普
万方数据