A Review of Offline Signature Verification Techniques

Among various biometric modalities, signature verification remains one of the most widely used methods to authenticate the identity of an individual. Signature verification, the most important component of behavioral biometrics, has attracted significant research attention over the last three decades. Despite extensive research, the problem still remains open to research due to the variety of challenges it offers. The high intra-class variations in signatures resulting from different physical or mental states of the signer, the differences that appear with aging and the visual similarity in case of skilled forgeries etc. are only a few of the challenges to name. This paper is intended to provide a review of the recent advancements in offline signature verification with a discussion on different types of forgeries, the features that have been investigated for this problem and the classifiers employed. The pros and cons of notable recent contributions to this problem have also been presented along with a discussion of potential future research The security requirements in today's world have placed biometrics at the center of an ongoing debate concerning its key role in a multitude of applications. Biometrics measures individuals' unique physical or behavioral characteristics with the aim of recognizing or authenticating the claimed identity. Physical biometrics includes modalities like fingerprints, retina, iris, DNA and facial patterns etc. Behavioral biometrics, on the other hand, exploits the behavioral characteristics of an individual like signature, voice, keystroke pattern or gait etc. to determine the identity. These diverse biometric modalities have received significant research attention of the pattern classification community over the last three decades and mature verification/authentication systems are available for modalities like face, iris, voice and signature etc. Among these diverse biometric verification modes, signature verification is undoubtedly the most wide used and accepted attribute for identity verification and is also the subject of our study. Despite significant research, the problem of signature verification remains open due to the wide diversity of challenges it offers. This paper is intended to provide a review of the recent signature verification techniques proposed in the literature along with the pros and cons of each and a comparative overview in terms of performance. The paper also summarizes the types of forgeries and the general steps involved in verification of signatures. Handwritten signature verification is simple, secure, cheap and acceptable all over the world. It is frequently employed to approve the transfer of resources of millions of people in the form of bank checks, credit card payments and other financial documents. Other official and legal documents requiring signatures can also be validated using signature verification techniques (15). Signature verification, like all other pattern classification problems, is typically categorized into traditional phases of preprocessing, feature extraction and classification. Among different problem scenarios offered by signature verification, discriminating a sample of genuine signature from a skilled forgery is known to be the most challenging task. This paper reviews the signature verification problem from different perspectives. We first present the categories of signature verification from the view point of data acquisition followed by a discussion on the common types of forgeries. We then present a general discussion on the phases of preprocessing, feature extraction and classification steps in a signature verification system followed by a review of some recent and significant research contributions to this problem. Finally, we conclude our discussion summarizing potential research directions on this subject.