Adversarial Attacks on Biometric Systems: Vulnerabilities and Defense Mechanisms in Facial and Fingerprint Recognition

Abstract

Biometric authentication systems are typically used for secure access to facilities, services or information through the use of physical characteristics (i.e., facial recognition and fingerprints). These types of systems are subject to being attacked through means of manipulating input data in order to gain unauthorized access. This paper explores several common attack techniques (e.g., spoofing, deepfakes and adversarial perturbations) and reviews existing defenses (such as liveness detection and multi-factor authentication) that have been created to counteract these attacks. In addition, the study will provide an overview of the current state of biometric systems regarding their effectiveness in preventing future attacks as well as identify gaps within the biometric security framework related to mitigating new threats.