On-line Signature Biometrics using Support Vector Machine

Handwritten signature has long been established as the most diffuse mean for personal verification in our daily life. It is commonly, in general, in all kind of legal documents and transactions. Therefore, within all biometric modalities, signature is probably the widest accepted. On the other hand, signature is a behavioural characteristic of individuals, and therefore is considered being weaker against fraud. Signature verification is the process in which, for a given signature who belongs to a user, a decision is made whether the signature has been made by that user, a genuine signature, or has been made by another user, a forgery signature. Typically, forged signatures have been classified into three groups: (1) random, (2) simple and (3) skilled [SB00]. The different methods for signature verification can be divided in two main groups: off-line (static) and on-line (dynamic). The off-line techniques are based in processing the digitalized grey-scale image of the signature written on a paper. On the contrary, on-line techniques take into account dynamic characteristics of the signature such as pressure exerted, tilts, position or velocity of the stylus. All this signals provide, not only information of the signature, but also information about the act of signing, which is consider more related to the specific user.