Distances#
Image-to-image distances that respect mass transport. wasserstein_distance()
solves a Beckmann (dynamic optimal-transport) problem; EMD wraps a
discrete Earth Mover’s Distance. Physical integration of a signal to a mass
requires a Geometry.
|
Unified access to Wasserstein distance computation between images with same mass. |
|
Write the output of the Wasserstein distance to a VTK file. |
|
Class to determine the EMD through cv2. |
Geometries#
|
Geometry of a physical domain, supporting geometric integration. |
|
Class containing information of a porous geometry. |
|
Geometry with weighted volume. |
|
One or two-dimensional geometry extruded to three dimensions. |
|
Class containing information of a porous geometry. |
Beckmann problem and solvers#
|
Base class for setting up the Beckmann problem. |
|
Class to determine the L1 EMD/Wasserstein distance solved with Newton's method. |
|
Class to determine the Wasserstein distance solved with the Bregman method. |
|
This contains the implementation of the GproxPDHG algorithm described in "SOLVING LARGE-SCALE OPTIMIZATION PROBLEMS WITH A CONVERGENCE RATE INDEPENDENT OF GRID SIZE" |
|
Mode for computing the l1 dissipation. |
|
Mode for computing the mobility. |
Advanced#
Class providing linear solver options for Beckmann's problem. |
|
Factory constructing a |
|
|
Enumeration of the linear-solver backends for the Beckmann problem. |
|
|
|
|
|
Direct solver for Beckmann's problem. |
|
|
|
Class to store and check the convergence criteria for the Beckmann problem. |
|
Class to store the convergence history. |