This toolbox reproduces the numerical results of the paper:

Justin Solomon, Fernando de Goes, Gabriel Peyré, Marco Cuturi, Adrian Butscher, Andy Nguyen, Tao Du, Leonidas Guibas, Convolutional Wasserstein Distances: Efficient Optimal Transportation on Geometric Domains, Proc. SIGGRAPH 2015.

The main directories are:

• data/: images and meshes datasets.
• code/: code directory, with the following sub-directories:
  • cpp/: C++ implementation of the algorithm.
  • figures/: Matlab scripts to reproduce the figure of the article.
  • tests/: Matlab scripts to reproduce some further examples not shown in the article.
  • convolutional_wasserstein/: Matlab main functions implementing the algorithms.
  • toolbox/: Matlab helper functions.
  • blur_functions/ and mesh_functions/: Matlab function to compute heat kernels.
  • colors_functions/: exernal library (c) Pascal Getreuer.
  • image_blur/: external library imgaussian (c) Dirk-Jan Kroon

Copyright (c) 2015, Justin Solomon, Fernando de Goes, Gabriel Peyré, Marco Cuturi, Adrian Butscher, Andy Nguyen, Tao Du, Leonidas Guibas