Advances In Digital Handwritten Signature Processing: A Human Artefact For E-society

A Human Artefact for e-Society

Giuseppe Pirlo, Donato Impedovo, Michael Fairhurst

• Publisher: WSPC
• Year: 2014
• eBook ISBN: 9789814579643
• Topic: Biological Sciences
• Subtopic: Science General
• Index: Biological Sciences

In the age of e-society, handwritten signature processing is an enabling technology in a multitude of fields in the “digital agenda” of many countries, ranging from e-health to e-commerce, from e-government to e-justice, from e-democracy to e-banking, and smart cities. Handwritten signatures are very complex signs; they are the result of an elaborate process that depends on the psychophysical state of the signer and the conditions under which the signature apposition process occurs. Notwithstanding, recent efforts from academies and industries now make possible the integration of signature-based technologies into other standard equipment to form complete solutions that are able to support the security requirements of today's society.

Advances in Digital Handwritten Signature Processing primarily provides an update on the most fascinating and valuable researches in the multifaceted field of handwritten signature analysis and processing. The chapters within also introduce and discuss critical aspects and precious opportunities related to the use of this technology, as well as highlight fundamental theoretical and applicative aspects of the field.

This book contains papers by well-recognized and active researchers and scientists, as well as by engineers and commercial managers working for large international companies in the field of signature-based systems for a wide range of applications and for the development of e-society.

This publication is devoted to both researchers and experts active in the field of biometrics and handwriting forensics, as well as professionals involved in the development of signature-based solutions for advanced applications in medicine, finance, commerce, banking, public and private administrations, etc. Handwritten Signature Processing may also be used as an advanced textbook by graduate students.

Contents:

• Stability Analysis of Online Signatures in the Generation Domain (Giuseppe Pirlo, Donato Impedovo, Rejean Plamondon and Christian O'Reilly)
• Exploiting Stability Regions for Online Signature Verification (Antonio Parziale and Angelo Marcelli)
• Two Bioinspired Methods for Dynamic Signatures Analysis (Jânio Canuto, Bernadette Dorizzi and Jogurta Montalvão)
• Using Global Features for Pre-Classification in Online Signature Verification Systems (Marianela Parodi and Juan C Gómez)
• Instance Selection Method in Multi-Expert System for Online Signature Verification (Giuseppe Pirlo, Donato Barbuzzi and Donato Impedovo)
• Towards a Shared Conceptualization for Automatic Signature Verification (Markus Liwicki, Muhammad Imran Malik and Charles Berger)
• Offline Signature Verification Based on Probabilistic Representation of Grid Events (Konstantina Barkoula, Elias N Zois, Evangelos Zervas and George Economou)
• Local Features for Off-Line Forensic Signature Verificaton (Muhammad Imran Malik, Markus Liwicki and Andreas Dengel)
• Emerging Issues for Static Handwritten Signature Biometrics (Moises Diaz-Caprera, Aythami Morales and Miguel A Ferrer)
• Biometric Signatures in Mobility: The Need for Transformation and the Opportunity for Innovation (Emilio Paterlini)
• Biometric Handwritten Solution: A World in a Signature (Carlo Nava)

Readership: Professionals, experts & researchers in the fields of biometrics and signature-based technology/solutions; advanced graduate students.

CHAPTER 1

STABILITY ANALYSIS OF ONLINE SIGNATURES IN THE GENERATION DOMAIN

G. Pirlo¹, D. Impedovo², R. Plamondon³, C. O'Reilly³

⁽¹⁾ Dipartimento di Informatica, Università degli Studi di Bari, Italy
{[email protected]}

⁽²⁾ Dip. Meccanica, Matematica e Management, Politecnico di Bari, Italy
{[email protected]}

⁽³⁾ École Polytechnique de Montréal, Canada
{rejean.plamondon, christian.oreilly}@polymtl.ca

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.

The experimental results were obtained using the SuSig database. They demonstrate that the new approach provides useful information on stability of online signatures and allows to better understand some characteristics of human behavior in signing.

1. Introduction

With the ubiquity of the internet in the modern society, the possibility to face actively social and economic transformations requires the development of new effective and efficient technologies and systems. Among others, the need for secure personal authentication is a crucial aspect in a multitude of applications related to the development of the Digital Agenda for e-health, e-government, e-justice, and so on.^1,2,3

For this purpose, biometrics has been considered with a renewed interest in recent years. Biometrics refers to individual recognition based on a person’s distinguishing characteristics. While other techniques use the possession of a token (i.e. badge, ID card, etc.) or the knowledge of something (i.e. a password, key phase, etc.) to perform personal recognition, biometric techniques offer the potential to use the inherent characteristics of the person to perform this task.³

Furthermore handwritten signature has a very special place in the wide set of biometric means that can be used for personal verification. Administrative and financial institutions recognize handwritten signatures as a legal means of verifying an individual’s identity. In addition, people are familiar with the use of signatures in their daily life. Therefore, it is not surprising that a special interest has been recently devoted to the field of automatic signature verification.^4,5

Unfortunately, a handwritten signature is a very complex trait. The rapid writing movement underlying handwritten signature generation is determined by a motor program stored into the signer’s brain and implemented through the signer’s writing system and writing devices (paper and pen type, etc.). Therefore, each handwritten signature strongly depends on a multitude of factors such as the psychophysical state of the signer and its social and cultural environment as well as the conditions under which the signature acquisition process occurs.^3,5 The result is that several basic aspects concerned with handwritten signature are still open to investigation. Among others, signature stability is currently at the centre of a large debate. In fact, everyone is aware that his/her signature is never the same even if each of us generally learns to sign at an early age and practices constantly to produce similar signatures according to his/her specific and personal model. Hence, stability in handwritten signatures is a crucial characteristic for investigating the intrinsic human properties related to handwriting generation processes concerning human psychology and biophysics. In addition, research on signature stability can provide new insights for a more accurate design of signature verification systems.⁵

In the scientific literature, approaches for the analysis of local stability in handwritten signature can be classified into three categories.

Approaches of the first category perform stability analyses on raw data. Examples of these approaches can be found with respect to both online and offline signatures. When online signatures are considered, a local stability function can be obtained by using Dynamic Time Warping (DTW) to match a genuine signature against other authentic specimens. Each matching is used to identify the Direct Matching Points (DMPs), that are unambiguously matched points of the genuine signature. Thus, a DMP can indicate the presence of a small stable region in the signature, since no significant distortion has been locally detected. The local stability of a signature point is determined as the average number of time it is a DMP, when the signature is matched against other genuine signatures. Following this procedure low- and high-stability regions can be identified^6,7,8 and usefully considered for selecting reference signatures^9,10 and implementing a verification strategy.^11,12 In another approach based on the handwriting generation and motor control studies, stability regions are defined as the longest common sequences of strokes between a pair of genuine signatures.¹³ When offline signatures are considered, the degree of stability of each signature region can be estimated by a multiple pattern-matching technique.^14,15 In this case, corresponding regions of genuine signatures are matched in order to estimate the extent to which they are locally similar. Of course, a preliminary step is used to determine the best alignment of the corresponding regions of signatures, in order to diminish any differences among them. Optical flow has also been used to estimate stability of offline signatures by investigating the shape deformation among genuine specimens^16,17 as well as considering the characteristics of the displacement vector fields.^18,19

Approaches of the second type use a signature model and perform the analysis of stability considering model parameters. For instance, a client-entropy measure has been proposed to group and characterize signatures in categories that can be related to signature variability and complexity. This measure based on local density estimated using a Hidden Markov Model (HMM) can be used to assess whether a signature contains or not enough information to be successfully processed by any verification system.²⁰

Approaches of the third category perform stability anal...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• Contents
• Preface
• Sponsors
• 1) STABILITY ANALYSIS OF ONLINE SIGNATURES IN THE GENERATION DOMAIN
• 2) EXPLOITING STABILITY REGIONS FOR ONLINE SIGNATURE VERIFICATION
• 3) TWO BIOINSPIRED METHODS FOR DYNAMIC SIGNATURES ANALYSIS
• 4) USING GLOBAL FEATURES FOR PRE-CLASSIFICATION IN ONLINE SIGNATURE VERIFICATION SYSTEMS
• 5) INSTANCE SELECTION METHOD IN MULTI-EXPERT SYSTEM FOR ONLINE SIGNATURE VERIFICATION
• 6) TOWARDS A SHARED CONCEPTUALIZATION FOR AUTOMATIC SIGNATURE VERIFICATION
• 7) OFFLINE SIGNATURE VERIFICATION BASED ON PROBABILISTIC REPRESENTATION OF GRID EVENTS
• 8) LOCAL FEATURES FOR OFF-LINE FORENSIC SIGNATURE VERIFICATON
• 9) EMERGING ISSUES FOR STATIC HANDWRITTEN SIGNATURE BIOMETRICS
• 10) BIOMETRIC SIGNATURES IN MOBILITY: THE NEED FOR TRANSFORMATION AND THE OPPORTUNITY FOR INNOVATION
• 11) BIOMETRIC HANDWRITTEN SOLUTION: A WORLD IN A SIGNATURE
• Index