mrfx

Old school discrete Markov Random Fields (and more) with JAX for regular lattices (images). Getting the most of JAX for the computational tasks

⚠️ Under heavy development. Check out the notebooks

Markov Random Fields

Unsupervised segmentation

Fast sampling

Discrete Gaussian Unitary Simplex

Fast sampling

Documentation

Contributing

• First fork the library.

• Then clone and install the library in development mode with

pip install -e .

• Install pre-commit and run it.

pip install pre-commit
pre-commit install

• Open a merge request once you are done with your changes.

Citing

@article{courbot2025gaussian,
  title={Gaussian Unit-simplex Markov random fields as a fast proxy for MRF sampling},
  author={Courbot, Jean-Baptiste and Gangloff, Hugo},
  year={2025}
}