Unbreakable Biometrics: How Physical Unclonable Functions are Revolutionizing Security

Abstract

Nowadays, technological advancements have pushed the growth of cyber-physical ecosystems through wearable and smart devices in the interoperability of the Internet of Things (IoT), cloud computing, and information technology. All these IoT devices are designed to launch multiple services or applications that require identity management systems, i.e., user access control systems for services, applications, resources, or data. Furthermore, access control systems are based on pattern recognition systems, where patterns cannot be forgotten, swapped, counterfeited, lost, or stolen. Consequently, biometric systems present an excellent opportunity area for security, social acceptance, performance, and experience quality for users, because biometric traits are unique to each individual and persist throughout their lifetime, preventing swapping, loss, or forgetting. Nonetheless, some biometric traits are not secrets, i.e., some traits can be acquired without an individual's knowledge and consent, e.g., a person's face or voice in social media videos. At the same time, the development of artificial intelligence has facilitated identity swap, attribute manipulation, and realistic expression exchange through image, audio, or video synthesis techniques known as DeepFakes.