An Improved D-LDA Based Face Recognition Using the Entire Information of Within-Class Scatter Matrix

摘要：

Direct linear discriminant analysis (DLDA) based methods have been successfully and widely used in face recognition, yet little work has been focused on the remaining subspace of within-class scatter matrix, in which exists more useful information for discrimination. In this paper we first present the traditional D-LDA based face recognition algorithm, and then calculate the remaining subspace by eliminating the spanning subspace of existing projecting vectors based on traditional D-LDA from subspace of Sw. Subsequently, adjusting factor is proposed, which is used to adjust the distances or projecting coordinates. The experimental results show that the modified D-LDA improves the recognition accuracy of a face recognition system.

作者: Fengqin Yu Yequan Huang Huizhong Yang

作者单位: School of Communication and Control Engineering Southern Yangtze University Wuxi, Jiangsu,China 214122

会议类型: 国际会议

会议名称: Fourth International Conference on Impulsive and Hybrid Dynamical Systems(ICIHDS 2007)(第四届国际脉冲和混合动力系统学术会议)

会议地点: 南宁

会议语种:英文

在线出版日期: 2007-07-20（万方平台首次上网日期，不代表论文的发表时间）