An online Chinese signature verification scheme based on the ART1 neural network

The authors propose an online Chinese signature verification method based on one-bit quantized pressure patterns which are the time domain information. The performance analysis was based on a set of ninety signatures from nine persons over a two-month period and ten traced forgeries obtained from ten forgers. Three experiments were analyzed in terms of type I and type II error probabilities. The results suggest that the timing information presented in one-bit pressure patterns can be used for the first stage screening of incoming signatures by using the ART1 neural network with multiple vigilance parameters.<>