Sparse and low-rank regularized deep subspace clustering

作者：

Highlights：

• Deep auto-encoder is employed to extract the deep features of images.

• The self-expression matrix is regularized using low-rankness to construct the affinity matrix.

• The sparse regularized deep representation of image is obtained to explore the discriminative information.

• The sub-gradient computation is devoted to solving the nuclear norm minimization problem.

摘要

•Deep auto-encoder is employed to extract the deep features of images.•The self-expression matrix is regularized using low-rankness to construct the affinity matrix.•The sparse regularized deep representation of image is obtained to explore the discriminative information.•The sub-gradient computation is devoted to solving the nuclear norm minimization problem.

论文关键词：Subspace clustering,Self-expressive matrix,Low-rank,Deep neural network

论文评审过程：Received 17 March 2020, Revised 27 June 2020, Accepted 28 June 2020, Available online 3 July 2020, Version of Record 4 July 2020.

论文官网地址：https://doi.org/10.1016/j.knosys.2020.106199