Person Recognition for Access Logging
Andrew Boka, Brendan Morris · 2019
The goal of this project is to develop a complete hardware and software system for access monitoring into a secured facility using modern facial recognition technology (FRT) and lightweight, inexpensive components. FRT detects and recognizes individuals as they enter and exit the secured area, providing accessible logs for a visualization of this activity. The project uses FaceNet FRT which takes a deep-learning approach to increase facial recognition accuracy. With input from two portable cameras, the FaceNet FRT is implemented on a Raspberry Pi using the Intel Movidius Neural Compute system.