AccessNet: A Three Layered Visual Based Access Authentication System for Restricted Zones

Abstract

Contactless authentication has become a trending topic these days, especially in this Corona pandemic situation, where contactless access systems are precautionary steps. The two most important contactless authentication systems right now are facial recognition and gait recognition. With having a possibility of spook attacks on facial recognition, gait patterns have gained popularity especially in the high-secured zones. Gait biometrics is considered as the next generation of biometrics which can enable identification of subjects even through a certain distance, avoiding any surface-contact. However, complex challenges gait patterns are mainly the view angles and distances for environmental setup and crucial gait features for algorithm setup. Hence, the authors are extending their previous work which is identifying ‘specific targets’ based on their gait patterns in Defence applications. We combined facial recognition designed with our anti-spoofing algorithm and gait patterns which gave on an average 92% accuracy irrespective of their background, motion, and angle.