State of the Art on Facial Recognition and Liveness Detection Technologies

Abstract

Facial recognition has become a crucial technology in biometric authentication systems, but it remains vulnerable to spoofing attempts, such as using fake photos or videos. As a result, liveness detection plays an important role differentiating genuine users from fraudulent ones. This work provides a state-of-the-art review of the main technologies in facial recognition and liveness detection. It first discuss recent deep learning-based advances in facial recognition, particularly those involving angular margin losses and embedding networks. Then, we analyze liveness detection approaches, distinguishing between passive methods (based on image analysis) and active methods (based on user interaction). This study examines their performance, benefits, limitations, and the latest trends and challenges in biometric security.