Canonical principal angles correlation analysis for two-view data.

Journal of Visual Communication and Image Representation(2016)

引用 11|浏览51
暂无评分
摘要
•Our method considers the correlation of two views, and exploits the local geometric structure of data.•The correlation of two views is measured by the angle between their principle components.•We introduce manifold regularization to maintain the local geometric.
更多
查看译文
关键词
Canonical correlation analysis,Feature fusion,Unpaired data,Manifold regularization,Mutual subspace method,Orthogonal projection,Pattern recognition,Handwritten recognition
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要