wooly_paths

Comment

Comment by Danny Andrés Osorio Gaviria on January 11, 2012 at 9:00pm

there are tutorials or something that maybe it can guide us to find good results like this??? :(

Comment by Nay Soe on October 11, 2010 at 1:36am

is it possible to see an example of how you guys are setting up the kangaroo file to do this?

Comment by David Reeves on August 9, 2010 at 11:48pm

!!
very cool.

My solution isn't quite as sophisticated. I started out with an inverse square summation attraction approach but I hit a wall computationally so I simplified to the closest point on closest curve strategy. It really let me bump up the thread count but the behavior doesn't seem as convincing. After taking a look at that paper though, I'd really like to give it another shot with some spring forces in action. Really interesting stuff.

Comment by Daniel Piker on August 9, 2010 at 4:07am

Nice work!
I've been trying the same thing recently too:
I modelled each thread as a chain of springs, with mutual attraction between all nodes (proportional to the inverse cube of the distance).
However I found that many threads with lots of subdivisions got quite heavy to try and do in real time with this approach because it's O(n^2).
Are you using a similar technique, or sorting for closest points first and only attracting to them ?
btw - have you seen Yiannis Chatzikonstantinou's very nice work along these lines?
http://blog.volatileprototypes.com/2010/06/force-directed-bundling-...
(That paper he links to on force-directed-edge-bundling is also pretty cool)