Modelling Geometrical Big Data Using Locally Refined B-Splines
Presentation given by Heidi Dahl (SINTEF) at the Geometrical Big Data Sciences micro-workshop at GDSPM15 Conference
Abstract. LRB-spline approximation can be considered the geometrical equivalent of jpeg compression, with a high compression rate in smooth areas. Its approximation of point clouds gives a good rate of compression for surfaces, and we expect the compression to be even more significant for approximations of volumes and higher-dimensional data. In this presentation we present results from the EU-funded projects VELaSSCo and IQmulus, using LRB-splines to model big data, e.g., terrain and simulation datasets from Particle Element Methods.