Manifold Discriminant Projection for face Recognition

摘要：

A novel manifold-based learning method named Manifold Discriminant Projection (MDP) is proposed. MDP not only preserves the samples local geometric structure, but also makes full uses of class-label information. Specifically, this method firstly uses the neighborhood relationship and class label information to classify the training sample set. For each training sample, there are two classes which are called neighbor class and non-neighbor class; Then, the inter-class scatter and intra-class scatter are defined for each training sample; Finally, the ratio of total inter-class scatter and total intra-class scatter is maximized to make the nearby samples with the same label are more compact, and nearby classes are better separated. Experiment results on ORL and FERET face databases verify the effectiveness of our proposed algorithm.

关键词： manifold discriminant projection dimensionality reduction feature extraction manifold learning face recognition

作者: Caikou Chen Yiming Yu Yu Hou

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: 345-348

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）