Impacto do tamanho de voxel e método de segmentação na análise morfológica de sapos do gênero Brachycephalus pela técnica de microCT

• Ano de defesa: 2019
• Autor(a) principal: Santos, Caio Márcio Sorrentino de Freitas Farias dos
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• 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 Nuclear; UFRJ
• Palavras-chave em Português: Microtomografia; Tomografia computadorizada por raios X; Processamento de imagens; Brachycephalus; CNPQ::ENGENHARIAS::ENGENHARIA NUCLEAR::APLICACOES DE RADIOISOTOPOS
• Link de acesso: http://hdl.handle.net/11422/13300

The Brazilian Rainforest inhabits many species of the Brachycephalus genre and, among them, the Brachycephalus ephippium species. X-ray computed tomography is a nondestructive imaging technique which allows for the visualization and analysis of internal microestructures of various samples. This technique was and still is applied on biological research for the study of these animals, enabling thorough description and characterization of new species. The species used in this study is of particular interest, since it was the first of this genre to be discovered, which makes it in a way the interspecies comparative basis. The cranium is of fundamental interest and it’s very used in microCT application works. Amid structures of interest in the cranium, some are of the order of the milimeter, sometimes less than half a milimeter. The effects of the voxel size of microCT images can therefore be relevant, and this is what was investigated in this work, both visually and quantitatively. Two well-known segmentation methods were used: Global (manual) and Otsu (automatic), and eight morphometric parameters were analyzed. There was significant variation due to voxel size alone, which led to information loss on the images for most evaluated parameters, while segmentation did not represent significant variation for most parameters, only for Tb.Pf. and Obj.N.