Deep Metric Learning-based Face Recognition Pipeline with Anti-spoofing on Raspberry-Pi Single-Board Computer

Abstract

Face recognition is an application of computer vision used for the identification of the persons present in an image / frame acquired through a camera sensor. Such a technology is already quite prevalent in multitude of cases such as security, biometric payments, augmented reality filters etc. In this paper, the concept, working and implementation of deep metric learning-based face recognition will be studied which is currently among state-of-the-art methods for face verification and recognition. Additionally, another module called anti-spoofing will be used to augment this face recognition pipeline to improve the security against photo and video based attacks that might be used by the malicious elements to spoof the face recognition system.