Scalable Person Identification System For Real-Time Applications

Abstract

Face recognition systems can play significant role in our every day lives. This paper proposes a scalable system for person identification based on face recognition methods and its implementation that utilizes queues, containers and microservices architecture. The proposed system uses a GPU acceleration therefore it can run in real-time. It utilizes two deep neural networks – Single Shot Multibox Detector (SSD) for a face detection and Facenet for a face recognition.