Regularized constraint subspace based method for image set classification

作者：

Highlights：

• Constraint subspace formed by partial common subspace may cause information loss.

• Combine the difference and orthogonal subspaces to form full rank constraint subspace.

• Generalize proposal to common framework by eigenspectrum regularization techniques.

• Propose a new eigenspectrum regularization model by concept of difference subspace.

摘要

•Constraint subspace formed by partial common subspace may cause information loss.•Combine the difference and orthogonal subspaces to form full rank constraint subspace.•Generalize proposal to common framework by eigenspectrum regularization techniques.•Propose a new eigenspectrum regularization model by concept of difference subspace.

论文关键词：Subspace method,Constraint subspace,Difference subspace,Orthogonal subspace,Eigenspectrum regularization model

论文评审过程：Received 5 April 2017, Revised 12 September 2017, Accepted 16 November 2017, Available online 20 November 2017, Version of Record 21 December 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.11.020