PLATAFORMA DE SERVICIOS DE RECONOCIMIENTO FACIAL PARA DETECCIÓN DE PRÓFUGOS DE LA JUSTICIA EN ECUADOR

Abstract

This document presents the design and implementation of a facial recognition platform conceived for government security organizations and the general public through web services. The main objective of the platform is to support security organizations in the processes of searching and capturing people who are required by justice, using facial recognition techniques applied into images generated from public and private sources.

The best results we obtained involved the use of the OpenCV deep learning based face detector, chosen in combination with a pre-trained facial recognition model based on deep metric learning. The testing was done using multiple video sources in different environments. The final result is a functional, scalable and reliable facial recognition platform, with a success rate of 92% and low response times.