A method for implicitly generating blue noise point sets. This method is based on the observation that volume preservation of a curly noise vector field and dithering can be understood as moving points along the streamlines of the vector field. It is shown that volume preservation keeps the points well separated when dithering is performed using a curly noise vector field. At the same time, this dithering greatly reduces the anisotropy produced by a regular lattice. Combining these properties, dithering using convolutional noise can effectively transform a regular lattice into a point set with blue noise properties. This implicit approach eliminates the need for prior computation of the point set. This makes our technique valuable when arbitrarily large point sets with blue noise properties are required. This method is compared to several other jitter-based methods as well as other methods for generating blue-noise point sets. Finally, several applications of curl noise dithering in 2D and 3D are shown.

References from:  https://people.compute.dtu.dk/jerf/papers/cnj.pdf

Modeling open cell aluminum foam (top) and copper foam (bottom). We used a cubic grid for the aluminum foam due to its higher porosity and an octahedral grid for the copper foam. The time in the corner is for a 1spp frame. The blue noise properties of our method helps it better avoid point clustering issues. The photos in the rightmost column, courtesy of Beihai Composite Materials (https://www.metalfoamweb.com/), provide some intuition on the appearance of real metal foams.