SetMargin loss applied to deep keystroke biometrics with circle packing interpretation

作者：

Highlights：

• A new keystroke identification method based on a novel loss function.

• We introduce the Circle Packing problem as a novel way to gain insights into learned feature spaces.

• Experiments on open-set keystroke biometric database with 78,000 subjects.

摘要

•A new keystroke identification method based on a novel loss function.•We introduce the Circle Packing problem as a novel way to gain insights into learned feature spaces.•Experiments on open-set keystroke biometric database with 78,000 subjects.

论文关键词：Keystroke biometrics,Circle packing,Deep learning,DML

论文评审过程：Received 23 April 2021, Revised 13 August 2021, Accepted 28 August 2021, Available online 29 August 2021, Version of Record 8 September 2021.

论文官网地址：https://doi.org/10.1016/j.patcog.2021.108283