Estrutura de Reconhecimento Facial com Raspberry PI para Segurança Inteligente Utilizando Técnicas de Aprendizagem de Máquina

Açucena de Gois Parente; Iago Magalhães de Mesquita; Sarah Frota Alves; Francisco Aldinei Pereira Aragão

doi:10.20906/CBA2022/3283

Açucena de Gois Parente Grupo IDSoft, Sobral, CE 62030-020
Iago Magalhães de Mesquita Eixo de Controle de Processo Industriais - Instituto Federal de Educação, Ciência e Tecnologia do Estado do Ceará (IFCE), Sobral 62042-030
Sarah Frota Alves Eixo de Controle de Processo Industriais - Instituto Federal de Educação, Ciência e Tecnologia do Estado do Ceará (IFCE), Sobral 62042-030
Francisco Aldinei Pereira Aragão Eixo de Controle de Processo Industriais - Instituto Federal de Educação, Ciência e Tecnologia do Estado do Ceará (IFCE), Sobral 62042-030

DOI: https://doi.org/10.20906/CBA2022/3283

Keywords: Electronic security, Facial biometry, Artificial intelligence

Abstract

This paper proposes an electronic security system of significant robustness, applying techniques of facial biometry. These systems are recurrent in the literature, and therefore have provided the advancement of artificial intelligence and computer vision techniques. Three techniques for facial attribute extraction and classification of images, both static and dynamic, with focus of the facial detection and recognition process were evaluated. The presented results demonstrate the relevance of the studied techniques, with accuracy above 98%. For field testing, a low-cost prototype was developed to suggest a functional and commercial solution.