Learning discriminative binary codes for finger vein recognition

作者：

Highlights：

• A novel discriminative binary code learning method is proposed.

• Maximizing inter-class scatter and entropy is proposed to make codes discriminative.

• Proposed codes consider relationship among subjects due to use of subject graph.

• Length of proposed binary codes is smaller due to small size of relation graph.

摘要

•A novel discriminative binary code learning method is proposed.•Maximizing inter-class scatter and entropy is proposed to make codes discriminative.•Proposed codes consider relationship among subjects due to use of subject graph.•Length of proposed binary codes is smaller due to small size of relation graph.

论文关键词：Biometric recognition,Finger vein recognition,Discriminative binary codes learning

论文评审过程：Received 13 July 2016, Revised 28 September 2016, Accepted 3 November 2016, Available online 9 November 2016, Version of Record 12 March 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.11.002