Automatic Signature Verification on Handheld Devices

Automatic Signature Verification on Handheld Devices

Marcos Martinez-Diaz (Universidad Autonoma de Madrid, Spain), Julian Fierrez (Universidad Autonoma de Madrid, Spain) and Javier Ortega-Garcia (Universidad Autonoma de Madrid, Spain)
DOI: 10.4018/978-1-60566-978-6.ch014


Automatic signature verification on handheld devices can be seen as a means to improve usability in consumer applications and a way to reduce costs in corporate environments. It can be easily integrated in touchscreen devices, for example, as a part of combined handwriting and keypad-based multimodal interfaces. In the last few decades, several approaches to the problem of signature verification have been proposed. However, most research has been carried out considering signatures captured with digitizing tables, in which the quality of the captured data is much higher than in handheld devices. Signature verification on handheld devices represents a new scenario both for researchers and vendors. In this chapter, we introduce automatic signature verification as a component of multimodal interfaces; we analyze the applications and challenges of signature verification and overview available resources and research directions. A case study is also given, in which a state-of-the-art signature verification system adapted to handheld devices is presented.
Chapter Preview


The current proliferation and ubiquity of electronic applications and services has motivated the need for user authentication means, which must be convenient and reliable at the same time. Nowadays, access control and user authentication are common tasks that are usually performed with tokens or passwords. In the case of handheld devices, user authentication is performed in most cases with passwords that are provided through keypad-based monomodal interfaces. Biometric-based systems represent an alternative to traditional user validation methods, as they rely on anatomical (e.g. fingerprint, iris) or behavioral (e.g. voice, signature) traits to authenticate a user (Jain, Ross & Pankanti, 2006). These traits cannot be in general forgotten or stolen without severe consequences for the user. As an example, it is now common to observe fingerprint verification systems in handheld and portable electronic devices (e.g. laptops). Due to the enhanced convenience provided to the users compared to the use of passwords, biometrics such as voice, fingerprint or signature can improve the usability of mobile user authentication applications.

Among biometric traits, signature is one of the most socially accepted as it has been used in financial and legal transactions for centuries. Despite its acceptance, automatic signature verification is a challenging task per se, as it must face at the same time the variability among genuine signatures from an individual (high intra-class variability) and the possibility of skilled forgers, which can imitate a signature with high accuracy (low inter-class variability). Moreover, forgers are not always possible to model during the design of a verification system due to their unknown nature. Consequently, reliable automatic signature verification is still an open issue, which can be corroborated by the notable variety of research works conducted on this subject in the last decades (Plamondon & Lorette, 1989; Fierrez & Ortega-Garcia, 2007).

Two main types of signature verification systems exist, depending on the nature of the information they use to perform verification. Off-line systems use static signature images, which may have been scanned or acquired using a camera. On-line or dynamic systems use captured signature time-functions, extracted from the pen motion. These functions can be obtained using digitizer tablets or touchscreen-enabled devices such as Tablet-PCs and smartphones. Traditionally, dynamic systems have presented a better performance than off-line systems as more information levels (e.g. pen speed, pen pressure, etc.) than the signature static image are available to perform verification (Plamondon & Lorette, 1989).

In this context, touchscreen-enabled handheld devices represent an appropriate computing platform for the deployment of dynamic signature verification systems as they provide both a pen-based input and reasonable computing power. These devices are already prepared for multimodal interaction, usually based on the combination of keypad, handwriting or speech modalities. As a matter of fact, commercial devices already provide handwritten character recognition as a text input alternative for years. Touchscreen-enabled smartphones have recently experienced an outstanding technological evolution, representing many new promising scenarios and applications for ubiquitous user interfaces and, specifically, for signature verification. They have gathered an increasing interest among the scientific and industrial communities as they provide multimodal capabilities (Oviatt & Lunsford, 2003) and a convenient way of interfacing with other systems; being thus able to host a wide range of user-centric applications (Ballagas, Borchers, Rohs, Sheridan, 2006). Verification of signatures or graphical passwords on handheld devices provides a convenient method for ubiquitous user authentication in commercial payments or financial transactions among other applications.

Complete Chapter List

Search this Book: