Selecting the Principal Feature Components in the Three-dimensional Parameter Space for Face Recognition

摘要：

This paper presents a novel face recognition method of selecting the principal feature components in the 3D parameter space constructed using the dimensions of three subspaces, I.e., PCA subspace, LDA subspace and LPP subspace, as axes. The global, local and clustering structure information can be used fully to enhance the recognition performance by selecting the principal feature component in 3D parameter space. The feasibility of the proposed method is successfully tested on the ORL and Yale face databases.

关键词： Face recognition principal component analysis linear discriminant analysis locality preserving projection three-dimensional parameter space

作者: Li Junbao Chu Shuchuan Pan Jengshyang

作者单位: Department of Automatic Test and Control,Harbin Institute of Technolog,Harbin,China Department of Information Management,Cheng Shiu University China Department of Electronic Engineering National Kaohsiung University of Applied Sciences,Kaohsiung,Tai

会议类型: 国际会议

会议名称: 第八届国际电子测量与仪器学术会议(Proceedings of 2007 8th International Conference on Electronic Measurement & Instruments)

会议地点: 西安

会议语种:英文

在线出版日期: 2007-08-16（万方平台首次上网日期，不代表论文的发表时间）