Biometric user authentication techniques have evoked an enormous interest by science, industry and society in the recent past. Scientists and developers have constantly pursued the technology for automated determination or confirmation of the identity of subjects based on measurements of physiological or behavioral traits of humans. Many biometric techniques have been implemented into authentication systems of apparently mature functionality. Although all biometric systems are subject to an intrinsic error-proneness, vendors argue for the accuracy of their systems by presenting statistical estimates or error probabilities, which suggest to be reasonably low for particular applications. With intentions like an expected integration of biometric features in travel documents in the near future, large-scale application of biometrics is becoming reality and consequently, an increasing number of people working in the IT domain will be confronted with biometric technology in the future. Thus, with the increasing diverseness of the different techniques, it is becoming more and more important for academic teachers, students and researchers as well as for practitioners, application designers and decision-makers in the IT security domain to understand the basic concepts, problems and limitations of authentication by biometrics. The goal and the uniqueness of this book are to impart knowledge to these people by expanding in two dimensions.
The horizontal dimension will educate the reader about the common principles in system design and evaluation of biometric systems in general, over the spectrum of the most relevant biometric techniques. Here, we will introduce to the state-of-the-art in biometric recognition across the variety of different modalities, including physiological traits such as fingerprint and iris recognition, as well as behavior such as voice or handwriting, and put them in context of the general theory of user authentication in IT security. For the system design part, we look at the systematic development of a generic process model, and provide an overview of implementation examples, along with references to selected signal and image processing methods, as well as classification techniques. The part on evaluation discusses the fundamental causes of inaccuracy of biometric systems and introduces mechanisms for experimental determination of quantitative figures in terms of error rates. Specifically for the most relevant biometrics, state-of-the-art recognition rates and references to technical details are presented for physiological methods based on face, fingerprint, iris, retina, ear and hand modalities, as well as for behavior-based approaches for voice, handwriting, gait, lip movement and keystroke dynamics. This comparative discussion will enable the reader to differentiate technical principles, potential applications and expected recognition accuracy across the zoo of different biometric schemes