Image decomposition based matrix regression with applications to robust face recognition

作者：

Highlights：

• LID is introduced in which using the local gradient distribution to decompose the image into several gradient images.

• The complex structure of the image is naturally separated and spread across gradient images.

• ID-NMR is proposed in which we combine the LID with NMR to model a simple and robust classifier.

• Experimental results on 6 datasets show that ID-NMR outperforms the state-of-the-art regression based classifiers.

摘要

•LID is introduced in which using the local gradient distribution to decompose the image into several gradient images.•The complex structure of the image is naturally separated and spread across gradient images.•ID-NMR is proposed in which we combine the LID with NMR to model a simple and robust classifier.•Experimental results on 6 datasets show that ID-NMR outperforms the state-of-the-art regression based classifiers.

论文关键词：Matrix regression,Low rank,Image decomposition,Pattern classification

论文评审过程：Received 10 May 2019, Revised 24 November 2019, Accepted 9 January 2020, Available online 5 February 2020, Version of Record 12 February 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107204