A fuzzy perception for off-line handwritten signature verification

A fuzzy perception for off-line handwritten signature verification

Simon, C. and Levrat, E. and Sabourin, R. and Bremont, J.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 1997

Abstract : The purpose of this paper is the assessment of a family of shape factors for off-line signature verification. The initial method, suggested by Sabourin [1], which extracts geometric features, is modified to assess the performance of verification in the general case of forgery. We contribute to the information coded by adapting the coding method to the image and by integrating a spatial distance into the shape factor definition. Moreover, we include the coding of an information related to the dynamic of the signature. To use these two types of information, we propose a fuzzy technique to combine and then obtain one kind of information. We evaluate this new coding operator with two types of forgery: random forgery and photocopy simulation with some adapted protocols.