Online signature verification using single-template matching with time-series averaging and gradient boosting

作者：

Highlights：

• We propose a novel single-template strategy for function-based online signature verification.

• The single-template strategy adopts a mean template set and multiple DTW distances.

• The mean template is created by a novel time-series averaging method, EB-DBA.

• The multiple DTW distances are combined using gradient boosting as the weighting scheme.

• The proposed method yields higher performance than existing signature verification systems.

摘要

•We propose a novel single-template strategy for function-based online signature verification.•The single-template strategy adopts a mean template set and multiple DTW distances.•The mean template is created by a novel time-series averaging method, EB-DBA.•The multiple DTW distances are combined using gradient boosting as the weighting scheme.•The proposed method yields higher performance than existing signature verification systems.

论文关键词：Biometrics,Forensics,Signature verification,Template matching,Variable importance,Gradient boosting,Dynamic time warping (DTW),Euclidean barycenter-based DTW barycenter averaging (EB-DBA)

论文评审过程：Received 10 July 2019, Revised 5 January 2020, Accepted 22 January 2020, Available online 27 January 2020, Version of Record 6 February 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2020.107227