Dicriminative Feature Weighted Hausdorff Distance for Face Recognition

Huachun Tan,Jianshuai Feng,Jie Zhang,Yujin Zhang,Wuhong Wang
DOI: https://doi.org/10.1016/j.egypro.2011.10.232
2011-01-01
Energy Procedia
Abstract:The different face regions have different degrees of importance for face recognition. In previous weighted Hausdorff distance (HD) measures for face recognition, the weighting function are proposed to try to describe the discriminative properties of face images effectively. Motivated by the basic idea of weighting function for HD, a new weighting function based on Linear Discriminant Analysis (LDA), due to its efficient discriminative ability, is proposed to weight different face edge points. The new weighted HD, termed Edge LDA Weighted HD (ELWHD), is applied to face recognition. Experiments comparing with pervious HD measures on ORL face database show the new method achieves higher recognition rate. (C) 2011 Published by Elsevier Ltd. Selection and/or peer-review under responsibility of the Organizers of 2011 International Conference on Energy and Environmental Science.
What problem does this paper attempt to address?