西安电子科技大学学报（自然科学版）2012，Vol.39Issue(5)：96-101,167,7.DOI:10.3969/j.issn.1001-2400.2012.05.017
类依赖增强线性判别分析算法
Classification method for multimodal data-class dependent ELDA
摘要
Abstract
Multimodal data refer to the data of a class that can be divided into two or more clusters. This paper proposes an improved method which is called the class-dependent and enhanced method. Enhanced LDA is combined with class-dependent LDA (CDLDA) to classify the multimodal data. In the new algorithm, we first use Enhanced LDA to reduce the dimensionality of multimodal data without losing the local structure and then get a projection matrix for each data class to obtain the characteristic differences for different data class distribution by the maximum scatter difference discriminant analysis criterion. Experiments on the face databases show the encouraging recognition performance of the proposed algorithm.关键词
多模态数据/分类/线性判别分析/增强线性判别分析/类依赖Key words
multimodal data/ classification/ LDA/ enhanced LDA/ class-dependence分类
信息技术与安全科学引用本文复制引用
任获荣,李春晓,孙建维,秦红波,何培培,高敏..类依赖增强线性判别分析算法[J].西安电子科技大学学报（自然科学版）,2012,39(5):96-101,167,7.基金项目
中央高校基本科研业务费资助项目(K50510040013) （K50510040013）
