On Monday, 19 May 2014 11:01:05 UTC+2, Simon Danisch wrote: > > @Job van der Zwan, can you elaborate on this? > Voronoi diagrams sound like a nice challenge for the API. > Delaunay-Triangulation and Voronoi diagrams are very important algorithms > for a lot of different use cases, like for example making a mesh out of > point clouds. > Seems like some algorithms are on their way: Fast, robust predicates with > Julia<https://groups.google.com/forum/#!searchin/julia-users/voronoi/julia-users/rpJs7O5e0pQ/pYE8rva4FqMJ> >
To be fair, I'm more interested in a 2D usecase than a 3D one, but I'm sure other people have a good use for a fast, accurate Voronoi library! In my case, I've been wanting to recreate Adrian Secord's work on using Weighted Voronoi diagrams as a way to generate stipple images[0], and then of course do further experiments with it (like continuing his ideas to decouple resolution from ink coverage). His work is over ten years old - by now I would have hoped computers would be fast enough to apply his algorithm in real-time to video-input! A library that would allow me to just feed a set of points at once, which would then return the Voronoi cells and their vertices would be ideal for this particular use-case, but I haven't found any (they mostly focus on immediately *drawing* the Voronoi map), and I'm not a good enough programmer to implement Fortune's algorithm[1] - it has me headscratching at weird bugs every time I've tried it. [0] http://cs.nyu.edu/~ajsecord/npar2002/html/index.html (PDF version: http://cs.nyu.edu/~ajsecord/npar2002/npar2002_ajsecord_preprint.pdf ) [1] http://en.wikipedia.org/wiki/Fortune%27s_algorithm
