随着计算机网络技术和生物识别技术的发展,通过生物特征进行身份认证成为信息时代的需要.由于长久以来作为授权标志,手写签名是最容易被大众接受的身份认证方式之一.本文改进了经典DTW中距离的计算方法,在考虑了时间序列自身形状特征的基础上,解决了在线手写签名中时间序列非线性弯曲的对正问题.经过签名认证实验表明,当ERR为3%时,签名认证的效果最好,进而证明在线手写签名适合作为身份认证手段.
With the development of computer network and biometric identification technology, identity verification with biologic characteristic feeds the need of the information age. Hand-written signature verification becomes a kind of the most acceptable way of identity verification because of it as the sign of authorization for a long time. In this paper, the distances measure of DTW is improved. Based on the characteristics of the shape of the time series itself, the problem that nonlinear warping of time series of on-line signature is solved. The experiment result is the ERR 3.0% when skilled forgeries are used. It is suitable to use On-line hand-written Signature as a mothod of identity verification.