Monte Carlo geometry processing: a grid-free approach to PDE-based methods on volumetric domains

Information

• Paper topic: Geometry
• Software type: Partial Code
• Able to run a replicability test: False
• Replicability score: 4
• Software language: c++
• License:
• Build mechanism: none
• Dependencies:
• Documentation score {0,1,2}: 2
• Reviewer: David Coeurjolly <david.coeurjolly@liris.cnrs.Fr>
• Time spent for the test (build->first run, timeout at 100min): 20min

Source code information

• Code URL: http://www.cs.cmu.edu/~kmcrane/Projects/MonteCarloGeometryProcessing/WoSLaplace2D.cpp.html
• Code URL 2: http://www.cs.cmu.edu/~kmcrane/Projects/MonteCarloGeometryProcessing/WoSPoisson2D.cpp.html
• MD5 hash of the code archive: 67657f1e1e644306f5c13877b5719694

Comments

The authors provide two 'quick and dirty/naive' c++ version of their Monte-Carlo PDE solver (Laplace and Poisson problems). The codes build and run perfectly outputing some CSV 2D map that must be converted to an image to 'visualize' the output. Per se, the provided code can only partially reproduce some paper results but interested readers should be able to quickly reimplement the core contributions of the paper (see Misc. comments). Some applications mentionned in the paper may require additional efforts (VF, Helmholtz Decomposition and Shape Deformation).

Misc. comments

In the project page, the authors also provide links to many shadertoy scripts that reproduce some paper results (scripts not from the authors).

If you want to contribute with another review, please follow these instructions.

Please consider to cut/paste/edit the raw JSON data attached to this paper.