Title: Dynamic signature recognition based on velocity changes of some features

Authors: Rafal Doroz, Piotr Porwik, Tomasz Para, Krzysztof Wrobel

Addresses: Institute of Informatics, University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland. ' Institute of Informatics, University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland. ' Institute of Informatics, University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland. ' Institute of Informatics, University of Silesia Bedzinska 39, 41-200 Sosnowiec, Poland

Abstract: Dynamic signature analysis allows us to register individuals and their hidden human behaviour. This paper presents a stroke-based approach to dynamic analysis of signature. Individual features can be identified by finding the discrete signature points like x,y-coordinates, pressure, time and pen velocity. Between signatures, the correlation measure is determined. The dynamic features are extracted from authentic and forged signatures. Experimental results show that measurement of dynamic features (velocity changes) contains important information and offers a high level of accuracy for signature verification in comparison with the results without such measurements, which will be explained in the following parts of the paper.

Keywords: biometrics; signature recognition; dynamic time warping; DTW; linear regression; dynamic measurements; feature measurement; signature analysis; signature verification; accuracy.

DOI: 10.1504/IJBM.2008.018663

International Journal of Biometrics, 2008 Vol.1 No.1, pp.47 - 62

Published online: 04 Jun 2008 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article