Over the past few decades, interest in algorithms for face recognition has been growing rapidly and has even surpassed human-level performance. Despite their accomplishments, their practical integration with a real-time performance-hungry system is not feasible due to high computational costs. So in this paper, we explore the recent, fast, and accurate face recognition system that can be easily integrated with real-time devices, and tested the algorithms on robot hardware platforms to confirm their robustness and speed.