Advances in Digital Handwritten Signature Processing

The following sections are included:

• Preface
• Contents
• SCIENTIFIC COMMITTEE
• SPONSORS

This paper presents a new approach for the analysis of local stability in online signature. Conversely to previous approaches in the literature, the analysis of stability is here performed by considering the characteristics of the processes underlying signature generation. For this purpose, the Sigma-Lognormal model developed in the context of the kinematic theory of rapid human movements is considered. It allows the representation of the information of the neuromuscular system involved in the production of complex movements like signatures…

We present a method for finding the stability regions within a set of genuine signatures and for selecting the most suitable ones to be used for online signature verification. The definition of stability region builds upon motor learning and adaptation in handwriting generation, while their selection exploits both their ability to model signing habits and their effectiveness in capturing distinctive features. The stability regions represent the core of a signature verification system whose performance is evaluated on a standard benchmark.

This work focuses on the problem of dynamic signature segmentation and representation. A brief review of segmentation techniques for online signatures and movement modelling is provided. Two dynamic signature segmentation/representation methods are proposed. These methods are based on psychophysical evidences that led to the well-known Minimum Jerk principle. These methods are alternatives to the existing techniques and are very simple to implement. Experimental evidence indicates that the Minimum Jerk is in fact a good choice for signature representation amongst the family of quadratic derivative cost functions.

In this paper, a pre-classification stage based on global features is incorporated to an online signature verification system for the purposes of improving its performance. The pre-classifier makes use of the discriminative power of some global features to discard (by declaring them as forgeries) those signatures for which the associated global features are far away from their respective means. Global features are considered individually and in a combined form for pre-classification. The subsequent classification stage is based on features obtained from a wavelet approximation of the time functions associated with the signing process. The experimental results show that the proposed pre-classification approach is capable of getting error rate improvements with respect to the case where no pre-classification is performed, having also the advantages of simplifying and speeding up the verification process.

In real world applications, signature verification systems should be able to learn continuously, as new signatures providing additional information become available. In fact, new data are not equally relevant for system improvement and a suitable data filtering strategy is generally required. In this context, instance selection is an important task for signature verification systems in order to select useful signatures to be considered for updating system knowledge, removing irrelevant and/or redundant instances from new data…

This chapter is an effort towards the development of a shared conceptualization regarding automatic signature verification especially between the Pattern Recognition (PR) and Forensic Handwriting Examiners (FHEs) communities. This is required because FHEs require state-of-the-art PR systems to incorporate them in forensic casework but so far most of these systems are not directly applicable to such environments. This is because of various differences, right from terminology to evaluation, in how the signature verification problem is approached in the two said communities. This chapter, therefore, addresses three major areas where the two communities differ and suggest possible solutions to their effect. First, it highlights how signature verification is taken differently in the above mentioned communities and why this gap is increasing. Various factors that widen this gap are discussed with reference to some of the recent signature verification studies and probable solutions are suggested. Second, it discusses the state-of-the-art evaluation and its problems as seen by FHEs. The real evaluation issues faced by FHEs, when trying to incorporate automatic signature verification systems in their routine casework, are presented. Third, it reports a standardized evaluation scheme capable of fulfilling the requirements of both PR researchers and FHEs and provides a practical exemplar of its usage.

A new grid based feature extraction methodology for offline handwritten signature representation is proposed which combines notions of grid feature extraction, set operations and information theory. It mutates the informative content of a set of binary mask elements (originally used to exploit the distribution formed by the transitions of signature pixels) by considering them as probabilistic events, simple or compound. Epigrammatically the feature definition procedure initiates by declaring an F_D-partition from a set of sixteen basic grid masks. In order to avoid exhaustive search among a huge number of possible F_D-partitions this work focuses on F_4SR-partitions of evenly distributed and orthogonal tetrads. All subset elements of an F₄-partition, considered as unique alphabet events that a discrete space source transmits when it is overlaid upon the signature image, are utilized as features and their distribution, as extracted by their mapping on the grid based binary signature image constitutes the novel proposed signature representation. Numerous ways of exploiting the entropy measure of the events distribution have been carried out in order to select the most appropriate scheme per writer. Results derived on two datasets indicate enhanced EER rates when compared to various state of the art approaches found in the literature.

The topic of signature verification has been addressed in the Pattern Recognition (PR) literature for several decades. The current PR research, at large, does not take the real needs of Forensic Handwriting Experts (FHEs) into account. The most crucial fact is that along with genuine and forged signatures, FHEs also have to deal with disguised signatures, i.e., when an author tries to make her/his own signatures look like a forgery. Our research in this area suggests that local or part-based features provide us a good possibility to identify disguising attempts. The purpose of this chapter is therefore twofold. First, we debate on the importance of disguised signatures with examples in order to draw the attention of PR community towards this important genre of signatures. Second, we present a novel comparison among three local features based off-line systems for forensic signature verification. The first system is a combination of scale-invariant Speeded Up Robust Features (SURF) and Fast Retina Keypoints (FREAK). The second system is based on a combination of Features from Accelerated Segment Test (FAST) and FREAK and the third system is based on nine local features with Gaussian Mixture Models (GMMs) classification. All of these systems are evaluated on the dataset of the 4NSigComp2010 signature verification competition which is the first publicly available dataset containing disguised signatures. Results indicate that our local features based systems outperform all the participants of the said competition both in terms of time and equal error rate.

This paper presents a review of the most recent advances in static/offline signature recognition using Computer Vision and also identifies some new trends and research opportunities such as the generation of synthetic signatures, time drifting, forger and disguise identification and multilingual scenarios. We conclude that the increasing collaboration between the Pattern Recognition community and Forensic Handwriting Experts will lead to future static handwriting signature milestones.

This chapter discusses how the growing demand for paperless processes in enterprise and government is pushing the development and widespread uptake of innovative technologies, with a strong focus on biometric signatures and new consumer devices. Two case studies explain the benefits of using biometric signatures in a business context.

This paper presents some advances in the application of biometric signature in HP. In particular, the proposal of HP is to provide a biometric signature infrastructure in order to start the process of acquisition of the contracts that are concluded with the end customer in a “paperless” mode. For the purpose, specific technical and legal rules have been addressed in order to offer complete biometric signature infrastructure solutions to be integrated into customer applications.

The following section is included:

• INDEX