Hi Magnus, yes I saw that paper recently and it looks interesting.
The process that it describes though is a pretty complex procedure, I think it would take really a considerable amount of effort to try and write all the code for that from scratch. The amount of time required would be significant enough that I don't think it would be feasible for me to attempt it anytime soon. Maybe if they had a library available to license that would be a possibility.
But also one other major thing - that paper does not really have anything to do with NURBS, it is about taking in triangle mesh data as input and producing quads as the output.
Because all the processing takes place in polygon data, that would actually be of the most benefit to have inside of a polygon modeling program rather than in MoI anyway, because then you would have the general ability to automatically re-topologize any kind of polygon data, not just ones that were generated from MoI.
So you may want to send that link over to the authors of whatever polygon modeling program you are currently using, since having it in that environment would actually give the most general benefit.
- Michael
|