multipers
Package for computing multiparameter persistence summaries. This is the code accompanying the article "Multiparameter persistence image for topological machine learning" (NeurIPS 2020), available here. The notebooks contain code for reproducing experiments (experiments.ipynb) and testing simple cases (visualization.ipynb).
The computation of multiparameter persistence images involves matching bars of different barcodes together. This can be done either by using your own function (e.g., the matching induced by the Wasserstein distance), or by using vineyards, see the documentation in the source code. In the latter case, you will have to compile the code in the dionysus_vineyards folder into a Python library, which you can easily do by running the first cell in experiments.ipynb or visualization.ipynb.
All files in the dionysus_vineyards folder (except for dionysus_vineyards.hpp, dionysus_vineyards.pyx and setup.py) are extracted from the Dionysus library, and can also be accessed here.
A custom implementation of the vineyards algorithm from the authors (which is also in C++) is also available upon request (by sending an email to mathieu.carriere3@gmail.com).