A Fusion Methodology for Recognition of Off-Line Signatures

In this paper we are presenting a work concerning the classification and recognition of off-line signatures. Signatures form a special class of handwriting in which legible letters or words may be impossible to exhibit but we can extract some features with the help of some parameters. Our proposed fusion methodology for improving the classification and recognition performance of classifiers is based on Dempster-Shafer evidence theory in which our contribution regarding to solve the problems like selection of focal elements and modeling the belief functions is also given. Distance classifiers studied, classify off-line signature images with the help of signature images projection along different axes and by employing some geometrical and fractal parameters which are explained in this article. Dempster-Shafer theory when applied for the fusion of these classifiers has improved the overall recognition rate.