In this talk, we propose the use of optimal transport for diffeomorphic registration of embedded surfaces and other sparse data. We also propose an extension to the space of images. This type of global similarity measures between data relies on the use of an embedding of data to a space of measures and the use of the entropic regularization of a generalization of the Wasserstein metric. This method can be generalized to many inverse problems.

This presentation is part of Minisymposium “MS28 - Diffeomorphic Image Registration: Numerics, Applications, and Theory (2 parts)”
organized by: Andreas Mang (Department of Mathematics, University of Houston) , George Biros (Institute for Computational Engineering and Sciences, University of Texas at Austin) .

Authors:

François-Xavier Vialard (University Paris-Dauphine)

Keywords:

image registration, inverse problems