作者: Augusto Salazar , Alexander Cerón , Flavio Prieto , None
DOI: 10.1007/978-3-642-17277-9_36
关键词:
摘要: The behavior of six curvature-based 3D shape descriptors which were computed on the surface face models, is studied. set includes k1, k2, Mean and Gaussian curvatures, Shape Index, Curvedness. Instead defining clusters vertices based value a given primitive feature, template composed by 28 anatomical regions, used to segment models extract location different landmarks fiducial points. Vertices are grouped by: themselves, region, region boundaries. aim this study analyze discriminant capacity each descriptor characterize regions identify key points facial surface. experiment testing with data from synthetic range images. In results: values, distributions, relevance indexes vertices, analyzed.