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)