Detecção e rastreamento de íris para implementação de uma interface homem-computador

• Autor(a) principal: Fernandes Junior, Valmir
• Data de Publicação: 2010
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do Mackenzie
• Texto Completo: http://dspace.mackenzie.br/handle/10899/24313
• topic: detecção e rastreamento de íris; interface homem-computador; problemas motores; iris detection and tracking; human-computer interface; mobile restrictions; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA

This paper presents a technique to iris detection and tracking that can be used in a human computer interface which allows people with mobility restricted, including no mobility above the shoulders, can control the mouse pointer only moving their eyes, without using expensive equipments. The unique data input used is an ordinary webcam without optical zoon, special lights or restricting user face mobility. The mouse displacement will occur in a straight way, in other words, the mouse cursor will be positioned at the place estimated by the technique. To the iris detection tests 60 images were used. 90.83% of the iris were identified correctly, there were 4.17% of missing iris and 5% false positives (iris were estimated in a wrong place). Using images generated straight from the webcam the iris were found correctly in 87,5%, no iris were found in 11,11% and in 1,39% the technique found iris in wrong places, the average time between positioning and a click is about 20 seconds.