Reconstrução tridimensional do corpo humano a partir de dados RGB-D

• Ano de defesa: 2014
• Autor(a) principal: Cuba, Manolo Canales
• Orientador(a): Gois, João Paulo
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal do ABC
• Programa de Pós-Graduação: Programa de Pós-Graduação em Ciência da Computação
• Link de acesso: http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=77312&midiaext=70677; http://biblioteca.ufabc.edu.br/index.php?codigo_sophia=77312&midiaext=70677/index.php?codigo_sophia=77312&midiaext=70676

The 3D reconstruction of human bodies is used in various areas, for example in medicine, lm, textile industry and the video game industry. Currently there are devices that make use of various technologies for obtaining three-dimensional data of the bodies. More accurate devices have high costs, making it impractical for use in domestic environments or small businesses, in addition to demand complex installations and knowledge of speci c software. Recent advances in the capture of three-dimensional data of objects and interiors have been provided, with applications like KinectFusion, which uses an RGB-D low-cost device. Although accurate for static scenes,these do not reconstruct with precision quasi-rigid objects such as human bodies, which have constant and involuntary movements such as breathing. This dissertation aims to study, develop and evaluate a reconstruction method of human bodies using a RGB-D device low cost, solving the problem of resolution and noise of the data on the device using super-resolution techniques and employing the merging of frames captured by the device, developing registration based on graphs of deformation.