Theory
------

See our algorithm description in our `computer vision paper <https://arxiv.org/pdf/2308.04526>`_.