Kernel two-dimensional ridge regression for subspace clustering
作者：
Highlights：
• Unlike existing methods that perform vectorization to 2D data in a pre-processing step, we propose to learn a 2D projection matrix such that the most expressive structural information is retained in the spanned subspaces.
• The learning of projection and construction of representation are seamlessly integrated, such that these two tasks mutually enhance each other and lead to powerful representation.
• Kernel method for 2D data is introduced to our model, which explicitly considers nonlinear structures of the data.
• Efficient optimization algorithm is developed with provable convergence guarantee.
• The algorithm does not rely on ALM type optimization as existing methods usually do, thus we do not need to introduce additional parameters in ALM framework.
• Extensive experiments confirm the effectiveness of our method.
摘要
•Unlike existing methods that perform vectorization to 2D data in a pre-processing step, we propose to learn a 2D projection matrix such that the most expressive structural information is retained in the spanned subspaces.•The learning of projection and construction of representation are seamlessly integrated, such that these two tasks mutually enhance each other and lead to powerful representation.•Kernel method for 2D data is introduced to our model, which explicitly considers nonlinear structures of the data.•Efficient optimization algorithm is developed with provable convergence guarantee.•The algorithm does not rely on ALM type optimization as existing methods usually do, thus we do not need to introduce additional parameters in ALM framework.•Extensive experiments confirm the effectiveness of our method.
论文关键词：Subspace clustering,Ridge regression,2-dimensional,Kernel
论文评审过程：Received 1 April 2020, Revised 24 October 2020, Accepted 1 November 2020, Available online 5 November 2020, Version of Record 19 February 2021.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107749
