Effect of Multidimensional Delay Vector in On-Line Handwriting Verification

Recently dynamic biometric technologies such as speech, online signature, gait pattern, are rapidly progressing. In order to evaluate the dynamic features from a time series, we proposed a novel similarity measurement, translation error. Translation error is computed by transition of multidimensional delay vector generated from an observed time series. In this work, we report the effect of multidimensional delay vector for online handwriting verification. Simulation results show that the multidimensional delay vector is a significant factor for improvement of verification accuracy.