Evaluation of Local and Global Features for Offline Signature Verification

In this paper we evaluate the impact of two stateof-the-art offline signature verification systems which are based on local and global features, respectively. It is important to take into account the real world needs of Forensic Handwriting Examiners (FHEs). In forensic scenarios, the FHEs have to make decisions not only about forged and genuine signatures but also about disguised signatures, i.e., signatures where the authentic author deliberately tries to hide his/her identity with the purpose of denial at a later stage. The disguised signatures play an important role in real forensic cases but are usually neglected in recent literaure. This is the novelty of our study and the topic of this paper, i.e., investigating the performance of automated systems on disguised signatures. Two robust offline signature verification systems are slightly improved and evaluated on publicly available data sets from previous signature verification competitions. The ICDAR 2009 offline signature verification competition dataset and the ICFHR 2010 4NSigComp signatures dataset. In our experiments we observed that global features are capable of providing good results if only a detection of genuine and forged signatures is needed. Local features, however, are much better suited to solve the forensic signature verification cases when disguised signatures are also involved. Noteworthy, the system based on local features could outperform all other participants at the ICFHR 4NSigComp 2010. Keywords-signature verification, mixture models, forgeries, disguised signatures, forensic handwriting analysis