Biometric-Authentication-of-Devices-and-Software

A comprehensive research report on the security and privacy aspects of biometric authentication systems, focusing primarily on facial, iris, and fingerprint recognition technologies. Authored by a team from the University of Central Florida, the report presents a detailed survey of existing literature and research in the field. It covers several key areas related to biometric authentication security, including facial authentication attacks such as data simulation attacks using the Fast Gradient Sign Method (FGSM), which can significantly reduce the accuracy of machine learning-based facial recognition systems. The report also discusses data poisoning attacks, where attackers infiltrate and sabotage the model's training dataset, and proposes a countermeasure called DEFEAT. Media extraction attacks, where content from social media is used to bypass authentication systems, are examined, along with EcoFace, an acoustic sensor-based detection method for these attacks. The authors propose a multi-modal authentication system combining facial, periocular, and iris recognition for enhanced security, using various feature extraction and comparison subsystems to improve accuracy and reduce false acceptance rates. The report also addresses iris authentication vulnerabilities, particularly print attacks, where adversaries use printed versions of irises to fool authentication systems. The authors provide a thorough analysis of these attacks and proposed countermeasures, offering insights into the current state of biometric authentication security. They highlight the ongoing challenges in securing these systems against sophisticated attacks and emphasize the need for continuous improvement in biometric authentication technologies. The report concludes that while biometric authentication offers significant advantages over traditional methods like PINs and passwords, it is not immune to security threats. The authors stress the importance of developing robust, multi-factor authentication systems that can withstand various types of attacks and ensure the privacy and security of users' biometric data. Overall, the report serves as a valuable resource for understanding the current landscape of biometric authentication security and the potential directions for future research and development in this field.