Skip to main navigation Skip to search Skip to main content

Template construction for function-based signature verification

  • Huimin Jiao*
  • , Dangxiao Wang
  • , Yuru Zhang
  • *Corresponding author for this work
  • Beihang University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

In function-based signature verification system, a robust template is critical to achieve high performance. In this paper, a two-staged selection method for template construction is proposed for function-based approach. At the first stage, the representative reference is selected for signature training before template constructed. From this reference, an initial template is picked up and the final template is created by averaging the references at the next stage. Experiments of the proposed methods are carried out and compared with other template construction methods. The result shows that the improved template typically represents the variability of intra-class. It can be pointed out that the proposed method reduces the false rejection rate remarkably, although a relatively little change happened to the false acceptation rate.

Original languageEnglish
Title of host publicationProceedings of the 2011 6th IEEE Conference on Industrial Electronics and Applications, ICIEA 2011
Pages2432-2437
Number of pages6
DOIs
StatePublished - 2011
Event2011 6th IEEE Conference on Industrial Electronics and Applications, ICIEA 2011 - Beijing, China
Duration: 21 Jun 201123 Jun 2011

Publication series

NameProceedings of the 2011 6th IEEE Conference on Industrial Electronics and Applications, ICIEA 2011

Conference

Conference2011 6th IEEE Conference on Industrial Electronics and Applications, ICIEA 2011
Country/TerritoryChina
CityBeijing
Period21/06/1123/06/11

Keywords

  • digital pen
  • force signal
  • function-based
  • signature verification
  • template constrction

Fingerprint

Dive into the research topics of 'Template construction for function-based signature verification'. Together they form a unique fingerprint.

Cite this