Many remeshing techniques sample the input surface in a meaningful way and then triangulate the samples to produce an output triangulated mesh. One class of methods samples in a parametrization of the surface. Another class samples directly on the surface. These latter methods must have sufficient density of samples to achieve outputs that are homeomorphic to the input. In many datasets samples must be very dense even in some nearly planar regions due to small local feature size. We present an isotropic remeshing algorithm called kCVT that achieves topological correctness while sampling sparsely in all flat regions, regardless of local feature size. This is accomplished by segmenting the surface, remeshing the segmented subsurfaces individually and then stitching them back together. We show that kCVT produces quality meshes using fewer triangles than other methods. The output quality meshes are both homeomorphic and geometrically close to the input surface.
CITATION STYLE
Edwards, J., Wang, W., & Bajaj, C. (2013). Surface segmentation for improved remeshing. In Proceedings of the 21st International Meshing Roundtable, IMR 2012 (pp. 403–418). Kluwer Academic Publishers. https://doi.org/10.1007/978-3-642-33573-0_24
Mendeley helps you to discover research relevant for your work.