def checkCycle(image):
# check if number of cycles in the donut image after thinning is 1
skel = Skeleton(image)
skel.setThinningOutput(mode="constant")
thinned_blob = skel.skeletonStack
label_img, countObjects = ndimage.measurements.label(thinned_blob, structure=np.ones((3, 3, 3), dtype=bool))
assert countObjects == 1, "number of cycles in single donut is {}".format(countObjects)
def checkAlgorithmSameObjects(image):
# check if same objects after and before pruning
skel = Skeleton(image)
skel.setPrunedSkeletonOutput()
label_img, countObjectsn = ndimage.measurements.label(skel.outputStack, structure=np.ones((3, 3, 3), dtype=np.uint8))
label_img, countObjects = ndimage.measurements.label(image, structure=np.ones((3, 3, 3), dtype=np.uint8))
assert (countObjectsn == countObjects) or np.sum(label_img == 2) == 1
def checkSameObjects(image):
# check if number of objects are same in input and output of thinning
label_img, countObjects = ndimage.measurements.label(image, structure=ndimage.generate_binary_structure(image.ndim, 2))
skel = Skeleton(image)
skel.setThinningOutput(mode="constant")
thinned_blob = skel.skeletonStack
label_img, countObjectsn = ndimage.measurements.label(thinned_blob, structure=ndimage.generate_binary_structure(image.ndim, 2))
assert (countObjectsn == countObjects), ("number of objects in input "
"{} is different from output {}".format(countObjects, countObjectsn))
return thinned_blob
def get_thinnedRandomBlob():
# get random convex blob
xs = np.random.uniform(-1, 1, size=50)
ys = np.random.uniform(-1, 1, size=50)
zs = np.random.uniform(-1, 1, size=50)
xyzs = list(zip(xs, ys, zs))
hullz = ConvexHull(xyzs)
xf, yf, zf = np.mgrid[-1:1:100j, -1:1:10j, -1:1:10j]
blob = np.ones(xf.shape, dtype=bool)
for x, y, z, c in hullz.equations:
mask = (xf * x) + (yf * y) + (zf * z) - c < 0
blob[mask] = 0
blob = blob.astype(bool)
skel = Skeleton(blob)
skel.setThinningOutput()
thinned_blob = skel.skeletonStack
return thinned_blob
def get_single_voxel_lineNobranches():
# graph of no branches single line
sampleLine = get_single_voxel_line()
skel = Skeleton(sampleLine)
skel.setNetworkGraph(False)
return skel.graph
def get_disjoint_trees_no_cycle_3d():
# graph of two disjoint trees
crosPair = get_disjoint_crosses()
skel = Skeleton(crosPair)
skel.setNetworkGraph(False)
return skel.graph
def get_cycles_with_branches_protrude():
# graph of a cycle with branches
sampleImage = get_tiny_loop_with_branches()
skel = Skeleton(sampleImage)
skel.setNetworkGraph(False)
return skel.graph
def get_cycle_no_tree():
# graph of a cycle
donut = get_donut()
skel = Skeleton(donut)
skel.setNetworkGraph(True)
return skel.graph
import pickle
from skeleton.skeletonClass import Skeleton
"""
Program to run and save statistics after pruning
"""
path = input("enter root directory of the binary pngs")
skeleton = Skeleton(path)
skeleton.segmentStatsAfterPruning()
pickle.dump(skeleton.statsAfter, open("skeletonStas.p", "wb"))
def checkAlgorithmPreservesImage(image):
skel = Skeleton(image)
skel.setThinningOutput(mode="constant")
thinned_blob = skel.skeletonStack
assert np.array_equal(image, thinned_blob)
import pickle
from skeleton.skeletonClass import Skeleton
"""
Program to run and save statistics after pruning
"""
path = input("enter root directory of the binary pngs")
skeleton = Skeleton(path)
skeleton.segmentStatsAfterPruning()
pickle.dump(skeleton.statsAfter, open("skeletonStas.p", "wb"))