Global discriminative-based nonnegative spectral clustering
作者：
Highlights：
• We maximize the between-class scatter matrix.
• Meanwhile, we minimize the within-class scatter matrix.
• We integrate the geometrical structure and discriminative structure in a framework.
• A global discriminative regularization term added into the objective functions.
• The new algorithms can preserve the global geometrical and discriminative structure.
摘要
•We maximize the between-class scatter matrix.•Meanwhile, we minimize the within-class scatter matrix.•We integrate the geometrical structure and discriminative structure in a framework.•A global discriminative regularization term added into the objective functions.•The new algorithms can preserve the global geometrical and discriminative structure.
论文关键词：Spectral clustering,Nonnegative matrix factorization (NMF),Global discrimination information
论文评审过程：Received 20 December 2014, Revised 17 October 2015, Accepted 29 January 2016, Available online 8 February 2016, Version of Record 21 March 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.01.035
