I have been searching a solution for weeks and not very fruitful and thus want to reach to this community for help. Here is my case, I have a 3D field defined on irregular data points (similar to octree grid, sparse at large scale and dense at a region of interest). As far as I know, to extract iso-surface Marching cube is out of the way as it handles regular grid points only. Interpretation using Voronoi tessellation based method can redefine the 3D field on regular grid, but the interpretation is slow in order to have a smooth iso-surface. Is there such an algorithm that can extract iso-surface efficiently from irregular data points? Would be even better if such an algorithm is implemented in a library for quick testing. Thanks.