I. Introduction
Automatic recognition of signature, fingerprint, voice and face image is considered as a branch of pattern recognition and biometrics. It aims to identify people from their biometric characteristics [1]. Recognition systems based on biometric data are categorized into physiological (fingerprint, retina etc) and behavioral (handwritten signature, voice etc) biometrics [2]. Identification and verification are two different tasks in recognition. Identification determines which user presents a given parameter among a set of known users. However, verification determines if the given parameter is provided by a known user or is a forgery. Verification handwritten signature represents a process of authenticating the identity of a person as a form of behavioral biometrics. Signature verification can be divided into two main areas depending on the data acquisition method used: offline signature verification and online verification signature. Online systems rely on dynamic information of a signature extracted at the time the signature is made. Offline systems use scanned image of a signature. Online signatures have extra information for extraction such as time, pressure, pen up and down, elevation, etc.