Automatic detection and visualization of distinctive structures in 3D unsteady multi-fields

摘要: Current unsteady multi-field simulation data-sets consist of millions data-points. To efficiently reduce this enormous amount information, local statistical complexity was recently introduced as a method that identifies distinctive structures using concepts from information theory. Due to high computational costs so far limited 2D data. In paper we propose new strategy for the computation is substantially faster and allows more precise analysis. The bottleneck original division spatio-temporal configurations in field (light-cones) into different classes behavior. algorithm uses density-driven Voronoi tessellation task accurately captures distribution sparsely sampled high-dimensional space. efficient achieved algorithms graph ability detect regions 3D illustrated flow weather simulations.