Sampling Point-Set Implicits

Abstract

We present a novel approach for point-set implicit surface sampling that is able to rapidly distribute particles over the surface of 3D objects. Our methods benefit from the inner structure of a MPU implicit to obtain a near-optimal initial distribution, with higher densities in areas of higher complexity. The adequate number of particles to effectively sample the surface is determined almost automatically and in accordance to the surface characteristics. We also use the MPU information to obtain local surface complexity heuristics that are used in the simulation stage of the particle system. Shape modeling operations are directly supported and the redistribution of particles is very fast for local edits. We present performance results that show that our system is faster than other state-of-the art approaches for the same number of points.