计算机应用研究2016，Vol.33Issue(12)：3686-3690,3712,6.DOI:10.3969/j.issn.1001-3695.2016.12.037
基于半监督典型相关分析的多视图维数约简
Semi-supervised canonical correlation analysis based multi-view dimensionality reduction
摘要
Abstract
In order to efficiently reduce dimensionality in multi-view semi-supervised scenarios,this paper proposed semi-su-pervised canonical correlation analysis methods which used nonnegative low-rank graph to propagate labels.Global linear neigh-borhoods captured by nonnegative low-rank graph could utilize information from both direct and reachable indirect neighbors to preserve the global cluster structures,while the low-rank property retained a compressed representation of the graph.After esti-mating label information of unlabeled samples by label propagation algorithm,it constructed soft label matrices of all samples and probabilistic within-class scatter matrices in each view.Then,by maximizing the correlations between samples of the same class from cross views and minimizing within-class variations in the low-dimensional feature space of each view simultaneously, it enhanced discriminative power of features.Experimental results demonstrate that the proposed methods can achieve better recognition performances and robustness than existing related methods and are effective multi-view dimensionality reduction methods.关键词
典型相关分析/人脸识别/多视图/维数约简/标签传播/半监督Key words
canonical correlation analysis(CCA)/face recognition/multi-view/dimensionality reduction/label propaga-tion/semi-supervised分类
信息技术与安全科学引用本文复制引用
董西伟,杨茂保,张广顺..基于半监督典型相关分析的多视图维数约简[J].计算机应用研究,2016,33(12):3686-3690,3712,6.基金项目
国家自然科学基金资助项目（61462048）；九江学院科研项目（2014KJYB019，2014KJYB030，2015LGYB26）；江西省教育厅科学技术研究项目 （）
