Package for estimating geometric and functional mappings between multiple modalities of data. Important applications include mapping tissue scale atlases to micron and submicron scale cellular and transcriptional data.
- Data: particles, each with position and feature value (discrete feature space)
- Geometric Mapping: rigid+scale and non-rigid deformation (diffeomorphism)
- Functional Mapping: single distribution over target feature space associated to each feature value in source space
-
xmodmap.io: datatypes, reading and saving
-
xmodmap.optimizer: optimizing functions, default is pytorch LBFGS
-
xmodmap.deformation: deformation models and functions
-
xmodmap.distances: fixed point costs including varifold distance (matching term), regularizer for functional mapping (kl divergence), and regularizer for support estimation parameter
-
xmodmap.models: selected parameters and mappings to estimate
-
examples: examples of running the code
-
tests: unit tests
Geometric Mapping Only example:
- 2D to 2D
python examples/example_2D_single_celltype.py
- 3D to 3D
python examples/example_3D_single_BEIALE_MTL.py
Geometric and Functional Mappings
- 2D to 2D
python examples/example_2D_cross_ratToMouseAtlas.py
- 3D to 3D
python examples/example_3D_cross_B2ToB5.py
- Kaitlin Stouffer (kstouff4@jhmi.edu)
- Alain Trouvé (alain.trouve@ens-paris-saclay.fr)
- Benjamin Charlier (benjamin.charlier@umontpellier.fr)