Characteristic extraction of Chinese signature identification based on B-spline function and wavelet transform

This paper presents a hybrid method to extract stable signature characteristic. We first develop a group of 4th order B-spline wavelet based on the better properties of B-spline function. After applying the novel B-spline wavelet to each stroke of the signature through wavelet decomposition, we design a set of formulas to synthesize characteristic value of each stroke so as to obtain signature characteristic values which have many good performances such as rotation invariance, translation invariance and scale invariance. Furthermore, a proper classifier is developed which can be more effective than the traditional Euclidean distance. The simulation results show that this method has better stability and reliability. © 2011 Springer-Verlag Berlin Heidelberg.