Learning locality-constrained collaborative representation for robust face recognition

作者：

Highlights：

• A coding algorithm integrates the locality with the global similarity of data.

• A new formulation of local consistency derives from similar inputs has similar codes.

• Our algorithm has an analytical solution and does not involve local minima.

• Consider the tasks of modeling facial images with various corruption and occlusions.

摘要

Highlights•A coding algorithm integrates the locality with the global similarity of data.•A new formulation of local consistency derives from similar inputs has similar codes.•Our algorithm has an analytical solution and does not involve local minima.•Consider the tasks of modeling facial images with various corruption and occlusions.

论文关键词：Non-sparse representation,Sparse representation,Local consistency,ℓ2-minimization,Partial occlusions,Additive noise,Non-additive noise,Robustness

论文评审过程：Received 23 October 2012, Revised 6 September 2013, Accepted 18 March 2014, Available online 27 March 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.03.013