The Research of handwritten signatures

Abstract

Developed an approach to identify the authenticity of signatures based on the variance, Experimental dataset was from SVC2004 Database, we combined variance into Dynamic Time Warping algorithm to calculate the intra-class distance (between real signatures) and inter-class distance (between real-forged signatures). The results show that the former is far less than the later, so we draw the conclusion that the deviation between real signatures is smaller than the real-forged ones when people signature. The method in this paper is simple and efficient, It also has strong stability and good recognition rate.