Hi Magnus,
> It might not be worth all the hard work required.
Well, it certainly could be useful but it is hard to say in advance exactly how well it would work.
The authors of that paper do not seem to have released any actual running program code for people to test their method with any mesh input.
So that makes it hard to really judge how "bulletproof" that method is, whether it has problems with certain situations or certain kinds of models, stuff like that.
That tends to make it quite a big risk for me to dedicate several months to working on it, since that would mean taking time off of other feature areas to achieve that...
It looks like a good area to watch though, maybe if the authors would release some program code for testing or make a library available for licensing it could make things more feasible.
- Michael
|