Regularized Wasserstein Means Based on Variational Transportation

12/02/2018

∙

by   Liang Mi, et al.

∙

0

∙

share

We raise the problem of regularizing Wasserstein means and propose several terms tailored to tackle different problems. Our formulation is based on variational transportation to distribute a sparse discrete measure into the target domain without mass splitting. The resulting sparse representation well captures the desired property of the domain while maintaining a small reconstruction error. We demonstrate the scalability and robustness of our method with examples of domain adaptation and skeleton layout.