Development of a Sigma–Lognormal representation for on-line signatures

摘要

This paper proposes an on-line signature representation based on Sigma–Lognormal modeling. It briefly overviews the prior art published on signature modeling and on human movement analysis for handwriting. Then it presents the Sigma–Lognormal paradigm and gives the key steps for the development of a completely automatic parameter extractor for complex human movements. Results of its application on signatures from a proprietary database and from the SVC2004 database are reported and analyzed in regards of the curve fitting quality. Other possible applications and future works are also suggested.