A discriminated correlation classifier for face recognition

  • Authors:
  • Zhongkai Han;Chi Fang;Xiaoqing Ding

  • Affiliations:
  • Tsinghua University, Beijing, China;Tsinghua University, Beijing, China;Tsinghua University, Beijing, China

  • Venue:
  • Proceedings of the 2010 ACM Symposium on Applied Computing
  • Year:
  • 2010

Quantified Score

Hi-index 0.00

Visualization

Abstract

In this paper, a discriminated correlation classifier is proposed to improve the performance of the two-dimensional (2-D) face recognition algorithm. Until now, many methods have been proposed to address the problems encountered by face recognition system, such as small number problem, pose and illumination variation, etc. All these works are aiming at enhancing the performance of the face recognition system. However, as far as we know, few work are concerning about how to improve the classifier, whose performance directly determines the final recognition accuracy. So, in this paper, our motivation is to improve the performance of correlation classifier, which is widely used in face recognition problems, to improve the recognition accuracy. Inspired by a correlation filter design method called Minimal Average Correlation Energy(MACE) filter, we propose a novel classifier called Discriminated Semi-Normalized Correlation (DSNC) classifier using our discriminative learning method. Compared with the classical discriminative learning methods that need many intra-class samples and can only be applied on close set recognition problems, our method needs only one intra-class sample and can be performed on open set face recognition problem. The validity of our method is tested on two benchmark face database(FRGC2.0 and FERET), and a private face database(THFaceID).