A NEW METHOD FOR SIGNATURE VERIFICATION BASED ON PHYSIOLOGICAL CHARACTERISTICS OF HAND MUSCLES AND TENDONS

摘要

Signature authentication with static and dynamic features of signature has been studied for decades, in this paper a novel and new method based on estimating elasticity and viscoelasticity characteristics of the muscles and tendons of index nger of the right hand was presented and the angles between the nger knuckles were collected by data collection glove and the location of digital pen tip on sensitive pad is stored in computer too. With NMC model and writing required mathematical equations and inverse modeling, physiological characteristics of muscles and tendons of right hand were estimated by LMS criteria. This approach has been applied on 30 right-hand persons that of each individual 5 genuine signature and some ordinary forgers to counterfeit genuine signature only by seeing the shape of original signature. 93.4 forgery signatures could have been recognized from genuine and only 6.6 could not have been detected. For verication, we used 5-fold cross-validation, with mean of EER= 3.57 and standard deviation of EER= 0.736. Therefore, we identied the physiological viscoelasticity and elasticity of muscles and tendons of hand as a new biometric.