iso splitting reaction diffusion

Comment

Comment by David Stasiuk on May 8, 2015 at 5:39am

Hi Daniel...thanks! Yes, the implementation reaches stability, and adjusting the diffU, diffV, F and K parameters all seem to produce the correct results. I haven't done the same confirmation test you did for yours, assigning variable parameters across a grid...but for us this was really just a tool to make an organic, isotropic pattern. The iso portion was so that we could leave one portion planar. Then, using the creases of the iso-cut, we tied it to the loop subdivision with creases implementation I did:

Comment by Daniel González Abalde on May 8, 2015 at 5:18am

Congratulations David, it looks great!!

But I am tempted to ask you... The patterns reaches spatiotemporal stability? have you control over the type of pattern (dots, labyrinth, waves...) changing the parameters? Or always get the same kind of pattern? That happened to me at first, I can lend a hand if you need/want.

Comment by David Stasiuk on May 8, 2015 at 2:30am

David...I'd love to, but unfortunately we are running our Rhino 5 on a zoo license and haven't been able to get the WIP working. But I'm super curious to see it!

Thanks Riccardo! The Gray Scott reaction diffusion algorithm here is implemented on a mesh, with each vertex capturing the U and V chemical values. The iso cut evaluates each vertex for its U value, and then splits faces along edges where one edge vertex is below the iso threshold set by the user, and the other is above (it interpolates the cut point along each edge by the target iso value on a range between the vertex values). All resulting quads get triangulated, and there's an additional laplacian smoothing process run on the mesh afterwards, with the split creases getting special treatment (they smooth as a spline).Plankton just makes all these operations much more straightforward.

The reaction diffusion algorithm is actually really simple. I adapted my implementation from Andrew Duncan's code here: http://www.openprocessing.org/sketch/25845.

Comment by Riccardo Majewski on May 8, 2015 at 1:20am

Great!

Can you explain how you achieve this? ... just more or less....

Comment by David Rutten on May 7, 2015 at 1:28pm

Oo! Put it through the Rhino WIP quad mesher, I wonder how it deals with this.

Comment by Pieter Segeren on May 7, 2015 at 1:12pm

Niiice!