Multi-subregion based correlation filter bank for robust face recognition

作者：

Highlights：

• A feature extraction algorithm MS-CFB is proposed for robust face recognition.

• MS-CFB combines the benefits of global-based and local-based feature extraction.

• We optimize overall correlation outputs by the multi-subregion based filter bank.

• The unconstrained form of correlation filter bank is designed in spatial domain.

• Experiments show the effectiveness of MS-CFB on popular face databases.

摘要

Highlights•A feature extraction algorithm MS-CFB is proposed for robust face recognition.•MS-CFB combines the benefits of global-based and local-based feature extraction.•We optimize overall correlation outputs by the multi-subregion based filter bank.•The unconstrained form of correlation filter bank is designed in spatial domain.•Experiments show the effectiveness of MS-CFB on popular face databases.

论文关键词：Correlation filter bank,Feature extraction,Face recognition

论文评审过程：Received 18 July 2013, Revised 25 March 2014, Accepted 3 May 2014, Available online 17 May 2014.

论文官网地址：https://doi.org/10.1016/j.patcog.2014.05.004