基于全程信息和时间弯曲校正的签名认证

Personal identification by using the whole dynamic signing information under nonlinear time warping

利用Wintab接口技术实现了对手写签名全程信息的实时采集 .提出一种非线性局部寻优时间弯曲校正方法 ,解决了手写签名全程信息时间序列的时间轴弯曲现象 .该校正方法不仅实现了对信息序列不同局部的非等强度校正 ,而且很好地保持了序列的单调性和连续性 .利用该校正方法在测试签名样本时获得了 96 %的身份认证正确率 .

This paper presents a method based on the Wintab interface to collect the whole dynamic signing information on real time. A nonlinear local optimal time warping algorithm is put forward to solve the problem of time warping in information series. The time warping transformation not only uses different slope weighting to process different parts of information series, but also well maintains the monotonicity and continuity of the series. This method achieves 96?% correct rate of personal identification in experiment.