A Kernel View of Manifold Analysis for Face Images

摘要：

This paper presents a new kernel method to analyse thehuman face images lying on the low dimensional mani-fold.Physical variations such as pose and illumination aremapped to the sematic feature space using a kernel matrixand an affine matrix.In this kernel method,the local geom-etry of the image data is modelled as generative units.Theglobal metric information is also preserved.The kernel for-mulation enables the manifold to be extended to the Out-Of-Sample data points.This provides a powerful tool for non-linear dimensional reduction,associative image denoisingand image synthesis.Extensive experiments are performed.to illustrate the theory.

作者: Dong Huang Zhang Yi Xiaorong Pu

作者单位: School of Computer Science and Engineering University of Electronic Science and Technology of China Chengdu 610054,P. R.China.

会议类型: 国际会议

会议名称: 2008 3rd International Conference on Intelligent System and Knowledge Engineering(第三届智能系统与知识工程国际会议)(ISKE 2008)

会议地点: 厦门

会议语种:英文

页码: 650-655

在线出版日期: 2008-11-17（万方平台首次上网日期，不代表论文的发表时间）