Multiscale morphology based illumination normalization with enhanced local textures for face recognition

作者：

Highlights：

• A multiscale morphology based illumination normalization method is proposed.

• The proposed method is robust to various illumination conditions.

• Enhancing local texture information further improved the recognition performance.

• The proposed method showed improved performance on several databases.

摘要

•A multiscale morphology based illumination normalization method is proposed.•The proposed method is robust to various illumination conditions.•Enhancing local texture information further improved the recognition performance.•The proposed method showed improved performance on several databases.

论文关键词：Face recognition,Illumination normalization,Texture information,Quotient image

论文评审过程：Received 22 March 2016, Revised 9 May 2016, Accepted 17 June 2016, Available online 18 June 2016, Version of Record 25 June 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.06.039