centroidal voronoi stippling

some early tests in GH to stipple an image. llyod's algorithm is used to generate an even voronoi distribution on the image. The centers of the voronoi cells are measured against image brightness. They are sampled and partitioned using kmeans. Each branch of like cells is randomly populated using an additive number of points in relation to the brightness. Realizing this strategy wasn't effective enough we developed our stippled using poisson's disk for stipple seed points. This helped produce far better results by more evenly spacing out the points. You can see those outcomes in the following images.