Desenvolvimento de uma interface computacional natural para pessoas com deficiência motora baseada em visão computacional

• Ano de defesa: 2018
• Autor(a) principal: Sampaio, Gustavo Scalabrini
• Orientador(a): Marengoni, Mauricio
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Presbiteriana Mackenzie
• Palavras-chave em Português: interação humano-computador; visão computacional; rastreamento facial; pontos faciais; pessoas com deficiência
• Área do conhecimento CNPq: CNPQ::CIENCIAS EXATAS E DA TERRA; CNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO
• Link de acesso: http://dspace.mackenzie.br/handle/10899/24473

New interaction devices such as touchscreens and virtual reality googles has made human computer interaction more intuitive and natural. This work presents concepts and techniques used for the development of a system that allows users to interact with personal computer in a simple and e cient way through face movements. Computer vision techniques such as segmentation, facial detection and landmarks detection were used, as well as programming and mathematical techniques for the development of this system. It is a natural human-computer interface that uses face, eye and mouth movements to perform control functions such as mouse cursor movement and triggering of clicks and keys. Several tests have shown that the developed system has superior performance compared to similar systems, 100 fps processing performance, throughput of 1:20 bits=s for mouse cursor movements, and 1.15 keystrokes with the face per second, it easy to use, fast learning and it can be used in a large number of applications such as internet browsing, social networking and games. The developed system can be used by people with motor disabilities as an assistive technology, promoting social inclusion, as well as providing greater educational and professional opportunities for this public.