Wasserstein Barycentric Coordinates: Histogram Regression Using Optimal Transport

This article defines a new way to perform intuitive and geometrically faithful regressions on histogram-valued data. It leverages the theory of optimal transport, and in particular the definition of Wasserstein barycenters, to introduce for the first time the notion of barycentric coordinates for histograms. These coordinates take into account the underlying geometry of the ground space on which the histograms are defined, and are thus particularly meaningful for applications in graphics to shapes, color or material modification. Beside this abstract construction, we propose a fast numerical optimization scheme to solve this backward problem (finding the barycentric coordinates of a given histogram) with a low computational overhead with respect to the forward problem (computing the barycenter). This scheme relies on a backward algorithmic differentiation of the Sinkhorn algorithm which is used to optimize the entropic regularization of Wasserstein barycenters. We showcase an illustrative set of applications of these Wasserstein coordinates to various problems in computer graphics: shape approximation, BRDF acquisition and color editing.

[1]  M. V. D. Panne,et al.  Displacement Interpolation Using Lagrangian Mass Transport , 2011 .

[2]  Raif M. Rustamov,et al.  Barycentric Coordinates on Surfaces , 2010, Comput. Graph. Forum.

[3]  Gabriel Peyré,et al.  Iterative Bregman Projections for Regularized Transportation Problems , 2014, SIAM J. Sci. Comput..

[4]  Marco Cuturi,et al.  Sinkhorn Distances: Lightspeed Computation of Optimal Transport , 2013, NIPS.

[5]  Richard Sinkhorn.  A Relationship Between Arbitrary Positive Matrices and Doubly Stochastic Matrices , 1964 .

[6]  Lei Zhu,et al.  Optimal Mass Transport for Registration and Warping , 2004, International Journal of Computer Vision.

[7]  Leonidas J. Guibas,et al.  A metric for distributions with applications to image databases , 1998, Sixth International Conference on Computer Vision (IEEE Cat. No.98CH36271).

[8]  L. Bregman.  The relaxation method of finding the common point of convex sets and its application to the solution of problems in convex programming , 1967 .

[9]  Guillaume Carlier,et al.  Barycenters in the Wasserstein Space , 2011, SIAM J. Math. Anal..

[10]  Yann Brenier,et al.  A computational fluid mechanics solution to the Monge-Kantorovich mass transfer problem , 2000, Numerische Mathematik.

[11]  Julien Rabin,et al.  Regularization of transportation maps for color and contrast transfer , 2010, 2010 IEEE International Conference on Image Processing.

[12]  Leonidas J. Guibas,et al.  Wasserstein Propagation for Semi-Supervised Learning , 2014, ICML.

[13]  Quentin Mérigot,et al.  A Multiscale Approach to Optimal Transport , 2011, Comput. Graph. Forum.

[14]  Thomas A. Funkhouser,et al.  A benchmark for 3D mesh segmentation , 2009, ACM Trans. Graph..

[15]  Arthur Cayley,et al.  The Collected Mathematical Papers: On Monge's “Mémoire sur la théorie des déblais et des remblais” , 2009 .

[16]  Jirí Filip,et al.  Template‐Based Sampling of Anisotropic BRDFs , 2014, Comput. Graph. Forum.

[17]  W. Deming,et al.  On a Least Squares Adjustment of a Sampled Frequency Table When the Expected Marginal Totals are Known , 1940 .

[18]  Gabriel Peyré,et al.  Fast Dictionary Learning with a Smoothed Wasserstein Loss , 2016, AISTATS.

[19]  Arnaud Doucet,et al.  Fast Computation of Wasserstein Barycenters , 2013, ICML.

[20]  Volkan Cevher,et al.  WASP: Scalable Bayes via barycenters of subset posteriors , 2015, AISTATS.

[21]  Marco Cuturi,et al.  Principal Geodesic Analysis for Probability Measures under the Optimal Transport Metric , 2015, NIPS.

[22]  Julien Rabin,et al.  Convex Color Image Segmentation with Optimal Transport Distances , 2015, SSVM.

[23]  Jérémie Bigot,et al.  Consistent estimation of a population barycenter in the Wasserstein space , 2013 .

[24]  Sylvain Paris,et al.  Example-based video color grading , 2013, ACM Trans. Graph..

[25]  Wojciech Matusik,et al.  A data-driven reflectance model , 2003, ACM Trans. Graph..

[26]  C. Villani.  Topics in Optimal Transportation , 2003 .

[27]  Erik Reinhard,et al.  Color Transfer between Images , 2001, IEEE Computer Graphics and Applications.

[28]  Michael Lindenbaum,et al.  Nonnegative Matrix Factorization with Earth Mover's Distance metric , 2009, 2009 IEEE Conference on Computer Vision and Pattern Recognition.

[29]  Julien Rabin,et al.  Wasserstein Barycenter and Its Application to Texture Mixing , 2011, SSVM.

[30]  Leonidas J. Guibas,et al.  Earth mover's distances on discrete surfaces , 2014, ACM Trans. Graph..

[31]  Jérémie Bigot,et al.  Geodesic PCA in the Wasserstein space by Convex PCA , 2017 .

[32]  David J. Kriegman,et al.  Toward a perceptual space for gloss , 2009, TOGS.

[33]  Julien Rabin,et al.  Sliced and Radon Wasserstein Barycenters of Measures , 2014, Journal of Mathematical Imaging and Vision.

[34]  Leonidas J. Guibas,et al.  The Earth Mover's Distance as a Metric for Image Retrieval , 2000, International Journal of Computer Vision.

[35]  Matthew Brand,et al.  Charting a Manifold , 2002, NIPS.

[36]  Mark W. Schmidt,et al.  Optimizing Costly Functions with Simple Constraints: A Limited-Memory Projected Quasi-Newton Algorithm , 2009, AISTATS.

[37]  Richard D. Neidinger,et al.  Introduction to Automatic Differentiation and MATLAB Object-Oriented Programming , 2010, SIAM Rev..

[38]  François Pitié,et al.  Automated colour grading using colour distribution transfer , 2007, Comput. Vis. Image Underst..