Door Access System Using Face Recognition

Abstract

This is a security system to restrict access to rooms using Raspberry Pi as main controller and implementation of Histogram of Oriented Gradients (HOG) feature extraction method. Upon face detection, the system is able to do face recognition for person identification by matching face embedding of the person face with one that already stored in database. This system also presents a random head-pose challenge to identify a real person from an image and then activate a solenoid to unlock the door. In this experiment, if the person is not wearing any face attributes, this system is tested to have 100% successful detection and recognition rate at 60 cm distance. If the person is wearing mask attribute, this system is tested to have 70% successful rate for detection and 45% successful rate for recognition at 60 cm distance.