A new affine registration algorithm applied to off-line signature verification

Signature identification and verification is considered among the most popular biometric methods in the area of personal authentication. The proposed method is based on offline signature verification by a novel algorithm for affine registration of true and false signatures 2D point sets. In the novel algorithm, each point in true and false signature is considered as a complex number and from each point set, a polynomial with complex coefficients can be computed whose roots are the points in the given point set. Then a verification function is achieved based on a difference between the points which can be determined by an unknown rotation. In order to archive the rotation, two-step algorithm is employed. We first reduce the affine registration problem to a rigid registration problem, and the unknown rotation is then computed using the coefficients of these polynomials. The algorithms have shown promising results while dealing with random forgeries and simple forgeries, also it gives better recognition rates in a public signature database.