Diseño de un sistema de detección facial utilizando cámaras IP para el reconocimiento de individuos en la cercanía de residencias familiares

Abstract

This project was designed to develop a prototype of a facial detection system for the recognition of individuals in the vicinity of family residences, since home security becomes an important issue to consider.

This design was developed with home security in mind using agile software development methodologies, in addition to hardware devices. The module mainly requires a camera that is located outside the home, to view people who are hanging around the family residence, for which it will be connected to a Raspberry Pi 4, which will host the face detection.

In the facial feature extraction tests, it was found that the system needs a minimum of 600 photos to improve the training of the system, obtaining 91.1% of recognized records, verifying the improvement in recognition accuracy in the members of the home group, with these tests the system was evaluated in different scenarios such as: luminosity, distances and sensitivity, thus the behavior of the system was verified at the time of facial recognition outside the home