Raspberry Pi-Based Access Control Using Face Recognition

Abstract

The aim of the research is the proposal and experimental implementation of a solution based on Raspberry Pi that is to allow face recognition with subsequent user notification. Proposal of whole system discusses the possible negative factors that can endanger functionality of the system. Once the face is recognized or not recognized, the message is sent to the administrator and access is permitted or denied. This represents the elevated level of system hardening. The system in this research is built on histogram of oriented gradients (HOG) and neural network (NN). Testing of the software system solution has been also carried out and proves the usability with detection and recognitions rates above 90%. The implemented system may be used as a part of complex IoT (Internet of Things) security solution.