Search

The story of how biometric devices have evolved is interesting in its own right and, in this chapter, we examine the early devices and techniques, noting how and perhaps why, some succeeded and went on to further development while some did not. We go on to acknowledge the step change that the events of September 2001 brought, with a renewed focus upon border control and the links with law enforcement. This development was more significant than many might suppose, and heralded a significant change in attitudes across several areas. For the biometrics industry, it was the boost that many needed to drive further development and performance enhancements. In short, this chapter provides us with a pocket evolution of biometrics technology as we have come to understand it, using electronic capture devices and the power of computers, to provide previously unimagined functionality.

Historically, biometric functionality tended to be ‘bolted on’ to existing applications rather than being properly integrated at the time of the broader application design. While this situation has improved somewhat in recent times, application design remains an area of concern with respect to biometric identity verification functionality. There remain operational factors which are not always properly understood, or seamlessly integrated into broader designs. Similarly, the broad cross section of potential users, especially with respect to public sector applications, are sometimes not fully taken into account within the design of operational systems. This chapter discusses an end to end system design approach and stresses factors such as the registration process, the biometric matching threshold, equivalence of realised performance, multi-modal biometrics, user interfaces and more. The BANTAM application modelling language is reintroduced and much guidance provided as to application design. In addition, both hardware and software are covered in a comprehensive overview of integrating biometric functionality into practical, operational systems.

Recent history has not always been kind to the biometrics industry. This is hardly surprising however, given some of the completely unrealistic claims made by suppliers in the past. Indeed, even today, we here ridiculous statements being made about the efficacy of biometric technology from a practical perspective. From an operational perspective, there remain some common assumptions and misconceptions which should have been dispelled many years ago. It seems that there is a requirement for some straight talking in this context, and this chapter rises to the challenge accordingly. There also exist areas where our understanding is incomplete and for which further discussion and associated research may be required. Consequently, current thinking around concepts such as biometric constancy, privacy and uniqueness is challenged and discussed. Finally, the whole is placed into a proper context of real world, operational systems and ideas are offered for taking a new and different approach to the concept of biometric identity verification. This is particularly important as we enter the age of big data and federated identity.

We consider the widespread erosion of personal privacy and question the trade-off with perceived security, restating the requirement for data protection in this context. The introduction of biometric technology provides additional challenges, not just from the data management perspective, but especially with regard to the assumptions made around the efficacy of a biometric match transaction. We therefore consider both the flow of a biometric throughout a typical system and the unrealistic assumptions which are often made around the result of a biometric match transaction and the validity of an associated identity. We offer suggestions as to a better way of using biometrics in this context and challenge the wholesale erosion of personal privacy. In addition to the consideration of practical matters, such as template storage, we revisit the whole concept of identity, and our often unrealistic focus on the same. This is a complex area which remains, in general, poorly understood.

It is pertinent to revisit the fundamentals of biometric technology, in order to ensure a proper understanding of what is actually happening within a biometric matching transaction. The biometric template is discussed, with various approaches considered as well as the granularity of observable features. The biometric capture device is then considered, from both an operational and practical maintenance perspective, before moving on to the biometric matching algorithm. The distinction between scientific theory and practical operation is made, stressing that, in the latter case, there exist several variables which do not occur under laboratory conditions and yet will impact upon realised performance. The biometric matching threshold is considered and a new model, the Proximity Matching Method, is introduced accordingly. This approach holds the promise for a much improved performance. The question of functions and routines as exposed by the device manufacturer receives some attention, as does the subject of data management and data flow throughout the system. Finally, we consider technology standards and their applicability to biometrics, before restating the importance of equivalence.

Biometrics are not new and, in this chapter, we trace the development of the idea of using anatomical and/or behavioural traits as personal identity verifiers, from ancient times, through the nineteenth and twentieth centuries, to the present time. The pioneering work of Alphonse Bertillon, Francis Galton, Juan Vucetich and others transformed identity verification for law enforcement purposes, but it was the electronic age which perhaps brought the larger transformation in the form of automated identity verification and, with computers, the ability to search through large databases of biometrics. This chapter provides an overview of the different types of biometrics, noting the potential advantages and disadvantages of the various techniques, while also exploring the primary application areas, their particular requirements and how we have utilised the technology within them. The fundamental operational principles of biometric identity verification are also explained, providing a background to concepts which will be explored in subsequent chapters.

This controversial opinion article questions and criticises the way the biometric business has evolved. While some good work has been done, the last 15 years has seen the industry tread water on many issues. Future efforts must be based on frank and open dialogue, where the interests of society as a whole are put first.

In this chapter we shall focus upon the definition of a system and consider how identity verification has developed from simple physical access control systems to more sophisticated bespoke systems and, finally, identity verification as a service. Biometric identity verification has mirrored this general trend. An important part of this transition is echoed in networks which have developed from closed-loop wired networks using traditional protocols such as RS485 and RS422, through more sophisticated computer networks based upon Ethernet or token ring, and finally to the IP-based networks which may span continents. We also discuss applications and the impact, especially, of border control and law enforcement applications in bringing biometric technology to a much broader potential market. The fundamentals of the biometric registration process are explained in the context of systems administration and the additional factors which biometric technology introduces in this respect. We discuss system architectures and infrastructures and introduce the concept of application modelling via the Biometric and Token Technology Application Modelling Language (BANTAM). This chapter takes the concept of automated biometric identity verification and applies it to systems and applications.

Much of the literature on biometric system design (as opposed to algorithmic, hardware or test design) has focused on system error rates and scaling equations [1–5]. In this chapter, I would like to step back a bit and ask “What are the operational goals and requirements of the system?” It sounds obvious, but before one can design and implement a successful biometric system, one must have a crystal-clear view of the requirements and how the operational processes will fit within the existing business processes and structures [6, 7]. It is tempting to get carried away with the technology and think of all the things we could do with biometrics and related technologies, when really we should concentrate on what we need to do in order to satisfy the requirements at hand, identifying how and where biometrics will be used to support the overall objective.

We have covered much ground in introducing BANTAM and exploring ways in which it may be used within a typical program related situation. Naturally, every organization will have its own way of working, but within this framework BANTAM may offer substantial benefits when dealing with biometric or related technology. In fact, creative users will find that BANTAM may be used to good effect within a variety of technology related program situations.

In this chapter we shall examine the distinction between identification and verification, the management of biometric templates and the impact this has on systems design, and we shall also take a look at the question of established performance criteria and how this relates to actual system performance. These are all issues which are important to understand if we are contemplating the use of biometrics. Furthermore, they are factors which should be considered in advance of any detailed planning with respect to biometric systems design.

Before we delve too deeply into applications for biometrics, it would perhaps be useful to spend a little time looking at how personal identification systems have developed and what the traditional applications have been for related technologies such as cards and tokens.

The preceding chapters of this book have hopefully provided the reader with a good background into biometric technology, how it has evolved and where it might be going in the future. We have even looked briefly at application development and the integration of biometrics into other processes. For many organizations however, having discovered the technology and considered potential applications within their particular sphere of operation, the next logical step would be to conduct some sort of trial or pilot scheme in order to prove the concept prior to a more ambitious implementation. The problem for many will be that this is a relatively new technology and that consequently they may not have a high level of in-house expertise or experience on which to draw. Furthermore, they may or may not be too enthusiastic about sharing the detail of their future plans and aspirations in this context with third party organizations, at least not at this stage. So how on earth do we go about setting up, running and evaluating a biometric pilot system? Well, hopefully this book will be of value in this respect as this chapter will be devoted to precisely that subject.