Subspace clustering using a symmetric low-rank representation
作者：
Highlights：
• A symmetric constraint is integrated into the low-rankness of data representation.
• A proof for minimizing the nuclear-norm problem with a symmetric constraint is given.
• The angular of the principal directions of the low-rank representation is exploited.
• eLRRSC attempts to seek a low-rank representation using a closed form solution.
• It generalizes the LRR-based subspace clustering framework with two key steps.
摘要
•A symmetric constraint is integrated into the low-rankness of data representation.•A proof for minimizing the nuclear-norm problem with a symmetric constraint is given.•The angular of the principal directions of the low-rank representation is exploited.•eLRRSC attempts to seek a low-rank representation using a closed form solution.•It generalizes the LRR-based subspace clustering framework with two key steps.
论文关键词：Low-rank representation,Subspace clustering,Affinity matrix learning,Spectral clustering
论文评审过程：Received 18 November 2016, Revised 26 February 2017, Accepted 28 February 2017, Available online 1 March 2017, Version of Record 12 May 2017.
论文官网地址：https://doi.org/10.1016/j.knosys.2017.02.031
