Pronounced water-zed. A simple watershed and region agglomeration library for affinity graphs.
Based on the watershed implementation of Aleksandar Zlateski and Chandan Singh.
python setup.py install
Requires numpy and cython.
import waterz
import numpy as np
# affinities is a [3,depth,height,width] numpy array of float32
affinities = ...
thresholds = [0, 100, 200]
segmentations = waterz.agglomerate(affinities, thresholds)