Reconhecimento facial com técnicas de machine learning

• Autor(a) principal: Rozar, Jéssyca Luiz
• Data de Publicação: 2020
• Outros Autores: Francisco, Antonio Marcos
• Tipo de documento: Trabalho de conclusão de curso
• Idioma: por
• Título da fonte: Repositório Universitário da Ânima (RUNA)
• Texto Completo: https://repositorio.animaeducacao.com.br/handle/ANIMA/11061
• topic: Inteligência artificial; Aprendizado de máquina; Redes neurais; OpenCV

For human beings to recognize people's faces by certain facial features is a natural skill. However, implementing facial recognition capabilities on machines is not an easy task. The machine language, known as machine learning, for facial recognition involves many mathematical calculations and requires great processing power. With the evolution of algorithms and the increase in the processing capacity of computers, the development of machines equipped with facial recognition technology is already possible. This technology is currently being used in several sectors, such as security, health, government sectors and others. Thus, the present work aims to carry out a broader study on the functioning of machine learning. It will also be discussed the functioning of artificial neural networks, a technique developed to simulate the functioning of the human neural network itself. Finally, a comparison of facial recognition algorithms provided by the openCV library will be presented. In order to evaluate the performance between the Eigenface, Fisherface and LBPH algorithms, it was necessary to develop a Python application with the ability to identify faces using global representations of the facial image.