Facial Recognition Based Workplace Security System Using LBPH Algorithm

Abstract

Workplace security is a major issue in today's working environment. To address the security concerns computer vision and its applications provide facial recognition-based security solutions. Every day, humans execute facial recognition effortlessly and almost without any effort. The goal of this paper is to create a reliable facial recognition system. This article presents a real-time face recognition-based security system for home and offices. It has been developed using the Local Binary Pattern Histogram (LBPH) algorithm. LBPH based Facial Recognition model has been developed and tested for security at workplaces. Variations in features, poses and angles of human faces as well as illumination have been selected as key parameters during testing of recognition algorithms. This model successfully passed over the two conditions and achieved more than 85% of accuracy. This kind of system can be used for video surveillance for streets, crowded areas and airports. Due to its resilience in varied lighting situations, the result suggests that LBPH FR is the best as compared to Eigenface, Fisherface.