Discriminative multi-layer illumination-robust feature extraction for face recognition

作者：

Highlights：

• Propose multi-layer decomposition model to extract illumination invariant features.

• Introduce weighting mechanism to control the contribution of each layer feature.

• Propose a discriminative filter to remove the influence of detrimental features.

• An optimization algorithm is proposed to determine weight and filter simultaneously.

摘要

•Propose multi-layer decomposition model to extract illumination invariant features.•Introduce weighting mechanism to control the contribution of each layer feature.•Propose a discriminative filter to remove the influence of detrimental features.•An optimization algorithm is proposed to determine weight and filter simultaneously.

论文关键词：Face recognition,Decomposition technique,Illumination-invariant feature,Discriminant filter,Small-scale

论文评审过程：Received 30 July 2016, Revised 30 November 2016, Accepted 2 February 2017, Available online 10 February 2017, Version of Record 20 February 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.02.004