A study on deep convolutional neural networks for computer vision applications

• Ano de defesa: 2017
• Autor(a) principal: Estevão Filho, Roberto de Moura
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Universidade Federal do Rio de Janeiro; Brasil; Instituto Alberto Luiz Coimbra de Pós-Graduação e Pesquisa de Engenharia; Programa de Pós-Graduação em Engenharia Elétrica; UFRJ
• Palavras-chave em Português: Redes neurais; Inteligência artificial; Aprendizado computacional; Jogos; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA::MEDIDAS ELETRICAS, MAGNETICAS E ELETRONICAS INSTRUMENTACAO
• Link de acesso: http://hdl.handle.net/11422/6098

Presents an Artificial Cortex model for solving complex problems. In the framework of the cortex, highlighting the use of machine learning based on multiple artificial neural networks. The design of the model of the observation of the functioning of the brain as a set of functional areas that operate in the process of cognition and subsequently use the acquired knowledge in decision-making and execution of complex actions. To experiment, observe and analyze the efficiency of the Cortex is decided by the application in an Intelligent Tutor System based on Digital Games, considering the teaching-learning process a complex subject, subjective and requires dynamism in their implementation (range strategies for Make the process flexible and personalized). The proposed tools are delimited through exploratory research, generating knowledge for practical use in the solution of problems related to education, such as applied research. Finally, statistical resources are used in order to explain the reality of the phenomenon, pointing out significant results (significance of 5%) in the improvement of learning, comparing teaching techniques.