Python implementation of the examples shown in the "Probabilistic Watershed: Sampling all spanning forests for seeded segmentation and semi-supervised learning" published in NeurIPS 2019. The repository also includes in the Supp_material folder the slides and poster used in the conference.
Our paper is mainly theoretical. We do not aim to provide a code of the Probabilistic Watershed, but to make available the code of the toy examples. In the paper we show the equivalence between the Random Walker by Leo Grady and the Probabilistic Watershed. For a code of the Probabilistic Watershed we refer to the Random Walker implementation of Scikit.