Off-line Signature Verification Using HMM for Random, Simple and Skilled Forgeries Edson J. R. Justino 1 , Flávio Bortolozzi 1 , Robert Sabourin 1, 2 1 PUCPR - Pontifícia Universidade Católica do Paraná ,R. Imaculada Conceição, 1155 CEP:80215-901 - Curitiba - PR – Brazil - {justino, fborto}@ppgia.pucpr.br 2 ETS - Ecole de Technologie Supérieure, 1100, rue Notre-Dame Ouest - Montréal (Québec) H3C 1K3 – Canada – sabourin@gpa.etsmtl.ca Abstract The problem of signature verification is in theory a pattern recognition task used to discriminate two classes, original and forgery signatures. Even after many efforts in order to develop new verification techniques for static signature verification, the influence of the forgery types has not been extensively studied. This paper reports the contribution to signature verification considering different forgery types in an HMM framework. The experiments have shown that the error rates of the simple and random forgery signatures are very closed. This reflects the real applications in which the simple forgeries represent the principal fraudulent case. In addition, the experiments show promising results in skilled forgery verification by using simple static and pseudodinamic features. 1. Introduction In an off-line signature verification system, a signature is acquired as an image [2-4]. This image represents a personal style of human handwriting, extensively described by the graphometry [7]. In such a system the objective is to detect three types of forgeries, which are related to intra and inter-personal variability [3]. The first type, called random forgery, is usually represented by a signature sample that belongs to a different writer of the signature model (see Fig. 1b). The second one, called simple forgery, is represented by a signature sample with the same shape of the genuine writer’s name (see Fig. 1c). The last type is the skilled forgery, represented by a suitable imitation of the genuine signature model (see Fig. 1d). (a) (b) (c) (d) Fig. 1 Type of forgeries: (a) genuine signature; (b) random forgery; (c) simulated simple forgery; and (d) simulated skilled forgery. Every type of forgery requests a different recognition approach. Methods based on Static approach are usually used to identify random and simple forgeries. The reason is that these methods have shown to be more suitable to describe characteristics related to the signature shape. For