Video Monitoring System using Facial Recognition: A Facenet-based Approach

Reductions in installation and storage costs have increased the demand for security systems, including video surveillance and digital authentication. The video surveillance systems, when monitored by humans, are subject to errors and are challenging to scale. Authentication systems can validate someone using a password or a card from another user. Facial recognition algorithms can solve this fault by the traffic monitoring of known individuals or intruders as well as for individual biometric authentication. Hence, this paper evaluates the FaceNet approach using the Labeled Faces in the Wild benchmark, as well as evaluates a machine learning technique known as support vector machine (SVM) for the classification of embedding generated using FaceNet. The suggested approach also models a real-time facial recognition system combining FaceNet and SVM, reaching 90% of accuracy using a medium webcam.