Local Mean Based Supervised Discrimiant Projection Algorithm

摘要：

Unsupervised discriminant projection (UDP) is an effective unsupervised discriminant projection technique based on locality, but it does not adopt the class information of training data which can enhance the discriminability. To overcome the limitation of UDP, the paper develops an enhanced UDP algorithm, called local mean based supervised projection algorithm. The proposed method does not only employs more robust local mean to preserve the local structure of data but also make use of the class information of data, which result in the better separability of transformed samples. Experimental results on the ORL > FERET and AR face databases verify the effectiveness of the proposed algorithm.

关键词： local mean neighborhood preserving feature extraction face recognition

作者: Caikou Chen Yiming Yu Min Jiang Xinlong Ding

作者单位: Information Engineering College, Yangzhou University, Yangzhou, 225009, China

会议类型: 国际会议

会议名称: 2010 International Conference on Circuit and Signal Processing(2010年电路与信号处理国际会议 ICCSP 2010)

会议地点: 上海

会议语种:英文

页码: 360-363

在线出版日期: 2010-12-25（万方平台首次上网日期，不代表论文的发表时间）