Auto-weighted 2-dimensional maximum margin criterion

作者：

Highlights：

• The proposed method extracts features from 2-order data, i.e., image, directly.

• Updating the weight automatically to make it insensitive to initialization.

• The objective value serves as a reflex of the performance in classification task.

摘要

•The proposed method extracts features from 2-order data, i.e., image, directly.•Updating the weight automatically to make it insensitive to initialization.•The objective value serves as a reflex of the performance in classification task.

论文关键词：Supervised learning,Auto-weighted parameter,2-dimensional criterion,Dimensionality selection,Classification

论文评审过程：Received 21 July 2017, Revised 29 April 2018, Accepted 20 May 2018, Available online 22 May 2018, Version of Record 18 June 2018.

论文官网地址：https://doi.org/10.1016/j.patcog.2018.05.021