Agglomerative Clustering on RAG

We start from a region adjacency graph as initial graph and use agglomerative clustering on this graph/

Out:

(303, 384) (303, 384)

from __future__ import print_function

# numpy[]
import numpy

# skimage
import skimage.filters       # filters
import skimage.segmentation  # Superpixels
import skimage.data          # Data
import skimage.color         # rgb2Gray
import matplotlib

# pylab
import pylab                # Plotting


# nifty
import nifty.graph.rag      # RAG
import nifty.graph.agglo    # Agglomerative clustering
import nifty.segmentation

# load some image
img = skimage.data.coins()


# slic superpixels
overseg = skimage.segmentation.slic(img, n_segments=2000,
    compactness=0.1, sigma=1)

# make the Region adjacency graph (RAG)
rag = nifty.graph.rag.gridRag(overseg)


# compute edge strength
smoothed = skimage.filters.gaussian(img, 2.5)
edgeStrength = skimage.filters.sobel(smoothed)


# accumulate the mean edge value
# along the superpixel boundaries
# length of each boundary and
print(edgeStrength.shape, overseg.shape)
edge_features, node_features = nifty.graph.rag.accumulateMeanAndLength(
    rag, edgeStrength, [512,512],0)
meanEdgeStrength = edge_features[:,0]
edgeSizes = edge_features[:,1]
nodeSizes = node_features[:,1]


# cluster-policy
clusterPolicy = nifty.graph.agglo.edgeWeightedClusterPolicy(
    graph=rag, edgeIndicators=meanEdgeStrength,
    edgeSizes=edgeSizes, nodeSizes=nodeSizes,
    numberOfNodesStop=25, sizeRegularizer=0.2)

# run agglomerative clustering
agglomerativeClustering = nifty.graph.agglo.agglomerativeClustering(clusterPolicy)
agglomerativeClustering.run()
nodeSeg = agglomerativeClustering.result()

# convert graph segmentation
# to pixel segmentation
seg = nifty.graph.rag.projectScalarNodeDataToPixels(rag, nodeSeg)



# plot the results

# increase default figure size
a,b = pylab.rcParams['figure.figsize']
pylab.rcParams['figure.figsize'] = 1.5*a, 1.5*b

f = pylab.figure()
f.add_subplot(2, 2, 1)
pylab.imshow(img, cmap='gray')
pylab.title('Raw Data')


f.add_subplot(2, 2, 2)
bImg = nifty.segmentation.markBoundaries(img, overseg, color=(1,0,0))
pylab.imshow(bImg, cmap='gray')
pylab.title('Superpixels')

f.add_subplot(2, 2, 3)
pylab.imshow(edgeStrength, cmap='gray')
pylab.title('Edge Strength')


f.add_subplot(2, 2, 4)
bImg = nifty.segmentation.markBoundaries(img, seg, color=(1,0,0))
pylab.imshow(bImg)
pylab.title('Segmentation')

pylab.show()

Total running time of the script: ( 0 minutes 0.489 seconds)