Exponential sparsity preserving projection with applications to image recognition

作者：

Highlights：

• An exponential sparsity preserving projection is proposed for solving the small-sample-size problem arising in sparsity preserving projection.

• Two efficient numerical methods are proposed for solving the exponential eigenvalue problem based on the structural features of ESPP.

• Some image recognition experiments on several real-world image databases illustrate the outperformances of ESPP.

摘要

•An exponential sparsity preserving projection is proposed for solving the small-sample-size problem arising in sparsity preserving projection.•Two efficient numerical methods are proposed for solving the exponential eigenvalue problem based on the structural features of ESPP.•Some image recognition experiments on several real-world image databases illustrate the outperformances of ESPP.

论文关键词：Sparsity preserving projection,Dimensionality reduction,Small-sample-size problem,Matrix exponential,Image recognition

论文评审过程：Received 8 August 2019, Revised 20 March 2020, Accepted 29 March 2020, Available online 4 April 2020, Version of Record 28 April 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107357