Two algorithms for fast 2D node generation: application to RBF meshless discretization of diffusion problems and image halftoning

Mesh generation techniques for traditional mesh based numerical approaches such as FEM and FVM have now reached a good degree of maturity. There is no such an acknowledged background when dealing with node generation techniques for meshless numerical approaches, despite their theoretical simplicity and efficiency; furthermore node generation can be put in connection with some well-known image approximation techniques. Two node generation algorithms are here proposed and employed in the numerical solution of 2D steady state diffusion problems by means of a local Radial Basis Function (RBF) meshless method. Finally, such algorithms are also tested for greyscale image approximation through stippling.