Monte Carlo Geometry Processing! WE NEED THIS in Blender!

Guys please look at it!
This is absolutely crazy, i can not even think on how many ways this can be applied!

1 Like

That link shows an empty shader for me, are you sure it’s the right one?

You possibly have NoScript enabled? :slight_smile:

The page displays here; Ubuntu 18.04, Firefox 78.

1 Like

Can you describe some of the applications in blender? Saw a video with a few examples and from what I’ve read from other knowledgeable folks, it does seem both revolutionary and easy to implement, but still a bit difficult for me to understand. I think it, at least, helps with fluid simulations.

Some related materials that will hopefully encourage more discussion…

Here’s a video on it by two minute papers:

applications mentioned in video:

  • multiple importance sampling
  • edit mesh with curves without creating intermediate triangle mesh
  • denoising
  • Helmholtz-Hodge decomposition

Link to paper by Rohan Sawhnen & Keenan Crane with example images and more shadertoy links:

From abstract:

This paper explores how core problems in PDE-based geometry processing can be efficiently and reliably solved via grid-free Monte Carlo methods. Modern geometric algorithms often need to solve Poisson-like equations on geometrically intricate domains. Conventional methods most often mesh the domain, which is both challenging and expensive for geometry with fine details or imperfections (holes, self-intersections, etc. ). In contrast, grid-free Monte Carlo methods avoid mesh generation entirely, and instead just evaluate closest point queries.

And an informative twitter thread from one of the authors

this stood out:

“Especially for problems with super complicated geometry, not having to mesh the domain provides a massive reduction in real world end-to-end cost. For instance, this model takes 14 hours to mesh and solve w/ FEM, but less than 1 minute to preview with Monte Carlo”


“As an added bonus, we can do geometry processing directly on nonmanifold meshes, implicit surfaces, instanced geometry, CSG, Bezier curves, NURBS surfaces, etc., without doing any tessellation, sampling, mesh booleans, etc”


“Finally, this all fits into existing geometry processing pipelines. Need a solution value at a vertex? Run our black-box solver. …Ok, if I say more I’ll probably just rewrite the paper here on Twitter! As you can tell, we’re really excited about the possibilities”