Название: Introduction to Biometrics, 2nd EditionАвтор: Anil K. Jain, Arun A. Ross, Karthik Nandakumar, Thomas Swearingen
Издательство: Springer
Серия: Texts in Computer Science
Год: 2025
Страниц: 418
Язык: английский
Формат: pdf (true)
Размер: 33.2 MB
This textbook introduces the fundamental concepts and techniques used in biometric recognition to students, practitioners, and non-experts in the field. Specifically, the book describes key methodologies used for sensing, feature extraction, and matching of commonly used biometric modalities such as fingerprint, face, iris, and voice. In addition, it presents techniques for fusion of biometric information to meet stringent accuracy requirements, also discussing various security issues and associated remedies involved in the deployment of biometric systems.
Biometric recognition, or simply biometrics, is the science and technology of establishing the identity of a person based on physical or behavioral attributes such as fingerprint, face, iris, and voice. By capturing these attributes using appropriately designed sensors, representing them in a digital format, and comparing this recorded data against the data acquired from the same person at an earlier time instance (enrollment), it is possible to automate the process of person recognition in real time. Thus, biometric recognition can be viewed as a Machine Learning problem, where the machine learns the salient features (patterns) in the biometric attributes of an individual and robustly matches such patterns efficiently and effectively with enrollment patterns.
Firstly, continuous improvements in sensing, processor and memory technologies have enabled the development of more compact and high fidelity sensors for biometric acquisition as well as real-time processing of biometric data for recognition of millions of persons. Secondly, advancements in the field of Machine Learning (especially deep neural network models) have revolutionized the field of biometrics by creating more salient representations (features) of biometric data and more accurate and robust biometric matchers or comparators. Finally, the advent of social media and smartphones has made it easier to collect large-scale multi-modal biometric datasets for training and benchmarking of some biometric modalities (e.g., face, voice). This is also accompanied by the possibility of generating large volumes of synthetic data through generative deep network models. This rapid transformation of the biometrics field has also increased concerns related to the security and privacy of biometric systems, including new threats such as deepfakes and inversion attacks on Machine Learning models and stored templates. At the same time, pain-staking standardization efforts from the biometrics community have borne fruit resulting in well-established standards and practices, which in turn has accelerated real-world deployment of biometric systems.
Furthermore, this second edition captures the progress made in the field of biometric recognition, with highlights including:
Lucid explanation of core biometric concepts (e.g., individuality and persistence), which builds a strong foundation for more in-depth study and research on biometrics
A new chapter on deep neural networks that provides a primer to recent advancements in Machine Learning and Computer Vision
Illustrative examples of how deep neural network models have contributed to the rapid evolution of biometrics in areas such as robust feature representation and synthetic biometric data generation
A new chapter on speaker recognition, which introduces the readers to person recognition based on the human voice characteristics
Presentation of emerging security threats such as deepfakes and adversarial attacks and sophisticated countermeasures such as presentation attack detection and template security
• Chapter 1 introduces the generic concepts in biometric recognition, including an overview of how a biometric system works, the terminology used to describe biometric traits and systems, the ways in which a biometric system can be used, the factors affecting the design of a biometric system, and the measures to evaluate the performance of a biometric system.
• Chapter 2 provides a primer on deep learning. It contains a basic overview of neural networks as well as a description of Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs). Examples of training and testing CNNs/RNNs are also provided using the Python programming language and the PyTorch package.
...
• Finally, the common security vulnerabilities of a biometric system and the various countermeasures that need to be followed to address these threats are presented in Chap. 9. In particular, this chapter emphasizes two of the most well-studied vulnerabilities of a biometric system, namely, presentation attack detection and biometric template protection (including matching in the encrypted domain).
While this textbook has been designed for senior undergraduate students and first-year graduate students studying a course on biometrics, it is also a useful reference guide for biometric system designers, developers, and integrators.
Contents:
Скачать Introduction to Biometrics, 2nd Edition