def test_tortuosity(self):
import skelet3d
data = np.zeros([20, 20, 20], dtype=np.int8)
# banana
data[5, 3:11, 4:6] = 1
data[5, 10:12, 5:12] = 1
# bar
data[5, 3:11, 15:17] = 1
# import sed3 as ped
# pe = ped.sed3(data)
# pe.show()
skel = skelet3d.skelet3d(data)
# pe = ped.sed3(skel)
# pe.show()
skan = sk.SkeletonAnalyser(copy.copy(skel),
volume_data=data,
voxelsize_mm=[1, 1, 1])
vessel_tree = skan.skeleton_analysis()
# banana
self.assertGreater(vessel_tree[1]['tortuosity'], 1.2)
# bar
self.assertLess(vessel_tree[2]['tortuosity'] - 1, 0.00001)
def test_skeleton_analyser_from_portal_vein_to_its_branches_and_save_to_pandas(self):
filename = "test_output_synthetic_porta.yaml"
# delete file if exists
if os.path.exists(filename):
os.remove(filename)
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = (
io3d.datasets.generate_synthetic_liver()
)
data_segm = segm == 2
# data_segm[41:44, 122:127, 68:70] = True
data_segm[40:45, 77:80, 100:110] = False
data_segm[42:44, 77:80, 103:106] = True
data_skelet = skelet3d.skelet3d(data_segm)
self.assertEqual(np.max(data_skelet), 1)
self.assertEqual(np.min(data_skelet), 0)
# import sed3
# ed = sed3.sed3(data_skelet, contour=data_segm)# , contour=branche_label)
# ed.show()
skan = sk.SkeletonAnalyser(copy.copy(data_skelet), volume_data=data_segm)
stats = skan.skeleton_analysis()
print(stats)
df = skan.stats_as_dataframe()
assert df is not None
def test_length(self):
data = np.zeros([20, 20, 20], dtype=np.int8)
data[2:13, 4, 4] = 1
data[6, 2:13, 6] = 1
data[8, 8, 2:13] = 1
# diagonala
data[11, 11, 11] = 1
data[12, 12, 12] = 1
data[13, 13, 13] = 1
# snake
# data[15:17, 13, 13] = 1
data[18, 14:17, 13] = 1
data[18, 17, 14:17] = 1
data[14:18, 17, 17] = 1
# import sed3
# ed = sed3.sed3(data)
# ed.show()
skel = data
skan = sk.SkeletonAnalyser(
copy.copy(skel),
volume_data=data,
voxelsize_mm=[1, 20, 300],
cut_wrong_skeleton=False,
)
# ed = sed3.sed3(skan.sklabel)
# ed.show()
vessel_tree = skan.skeleton_analysis()
self.assertAlmostEqual(vessel_tree[1]["lengthEstimationPixel"], 10)
self.assertAlmostEqual(vessel_tree[2]["lengthEstimationPixel"], 200)
self.assertAlmostEqual(vessel_tree[3]["lengthEstimationPixel"], 3000)
diag_length = 2 * ((1**2 + 20**2 + 300**2)**0.5)
self.assertAlmostEqual(vessel_tree[4]["lengthEstimationPixel"],
diag_length)
# test spline
self.assertLess(
vessel_tree[3]["lengthEstimationPixel"] -
vessel_tree[3]["lengthEstimationSpline"],
0.001,
)
# test poly
self.assertLess(
vessel_tree[1]["lengthEstimationPixel"] -
vessel_tree[1]["lengthEstimationPoly"],
0.001,
)
def test_fileter_small(self):
import skelet3d
data = np.zeros([20, 20, 20], dtype=np.int8)
data[5, 3:17, 5] = 1
# crossing
data[5, 12, 5:13] = 1
# vyrustek
data[5, 8, 5:9] = 1
data = skelet3d.skelet3d(data)
# pe = ped.sed3(data)
# pe.show()
skan = sk.SkeletonAnalyser(copy.copy(data))
output = skan.filter_small_objects(data, 3)
# skan.skeleton_analysis()
# pe = ped.sed3(output)
# pe.show()
self.assertEqual(output[5, 8, 7], 0)
def test_skeleton_analyser_from_portal_vein_to_its_branches(self):
filename = "test_output_synthetic_porta.yaml"
# delete file if exists
if os.path.exists(filename):
os.remove(filename)
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = (
io3d.datasets.generate_synthetic_liver())
data_segm = segm == 2
# data_segm[41:44, 122:127, 68:70] = True
data_segm[40:45, 77:80, 100:110] = False
data_segm[42:44, 77:80, 103:106] = True
data_skelet = skelet3d.skelet3d(data_segm)
self.assertEqual(np.max(data_skelet), 1)
self.assertEqual(np.min(data_skelet), 0)
# import sed3
# ed = sed3.sed3(data_skelet, contour=data_segm)# , contour=branche_label)
# ed.show()
skan = sk.SkeletonAnalyser(copy.copy(data_skelet),
volume_data=data_segm)
branche_label = skan.get_branch_label()
# import sed3
# ed = sed3.sed3(branche_label) #, contour=skan.sklabel)
# ed.show()
# import sed3
# ed = sed3.sed3(skan.sklabel)# , contour=branche_label)
# ed.show()
self.assertGreater(np.max(branche_label), 2)
# self.assertLess(np.max(branche_label), -4)
self.assertEqual(branche_label[0, 0, 0], 0)
def test_nodes_aggregation(self):
data = np.zeros([20, 20, 20], dtype=np.int8)
voxelsize_mm = [14, 10, 6]
# snake
# data[15:17, 13, 13] = 1
data[17, 3:17, 12] = 1
data[17, 17, 13:17] = 1
data[17, 9, 4:12] = 1
data[17, 14, 12:19] = 1
# data[18, 18, 15:17] = 1
# T-junction on the left
data[17, 4:16, 3] = 1
# import sed3
# ed = sed3.sed3(data)
# ed.show()
skel = data
skan = sk.SkeletonAnalyser(
copy.copy(skel),
volume_data=data,
voxelsize_mm=voxelsize_mm,
cut_wrong_skeleton=False,
aggregate_near_nodes_distance=20,
)
vessel_tree = skan.skeleton_analysis()
# ed = sed3.sed3(skan.sklabel, contour=data)
# ed.show()
# import ipdb; ipdb.set_trace() # noqa BREAKPOINT
# number of terminals + inner nodes
self.assertEqual(np.min(skan.sklabel), -7)
# number of edges
self.assertEqual(np.max(skan.sklabel), 6)
def test_skeleton_analyser_from_portal_vein(self):
filename = "test_output_synthetic_porta.yaml"
# delete file if exists
if os.path.exists(filename):
os.remove(filename)
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = io3d.datasets.generate_synthetic_liver(
)
data_segm = segm == 2
data_skelet = skelet3d.skelet3d(data_segm)
self.assertEqual(np.max(data_skelet), 1)
self.assertEqual(np.min(data_skelet), 0)
skan = sk.SkeletonAnalyser(copy.copy(data_skelet),
volume_data=data_segm)
skan.skeleton_analysis()
# self.assertEqual(output[5, 8, 7], 0)
skan.to_yaml(filename)
self.assertTrue(os.path.exists(filename))
def test_length_types(self):
"""
Test for comparation of various length estimation methods.
No strong assert here.
"""
data = np.zeros([20, 20, 20], dtype=np.int8)
# snake
data[8:10, 14, 13] = 1
data[10:12, 15, 13] = 1
data[12:14, 16, 13] = 1
# data[18, 14:17, 13] = 1
# data[18, 17, 14:17] = 1
# data[14:18, 17, 17] = 1
skel = data
skan = sk.SkeletonAnalyser(copy.copy(skel), volume_data=data,
voxelsize_mm=[1, 20, 300])
# skan.spline_smoothing = 5
vessel_tree = skan.skeleton_analysis()
pixel = vessel_tree[1]['lengthEstimationPixel']
poly = vessel_tree[1]['lengthEstimationPoly']
spline = vessel_tree[1]['lengthEstimationSpline']
def test_small_with_write_to_file(self):
filename = "test_output.yaml"
# delete file if exists
if os.path.exists(filename):
os.remove(filename)
data = np.zeros([60, 60, 60], dtype=np.int8)
data[5:8, 13:27, 5:9] = 1
# crossing
data[6:21, 18:20, 4:7] = 1
data_skelet = skelet3d.skelet3d(data)
# pe = ped.sed3(data)
# pe.show()
skan = sk.SkeletonAnalyser(copy.copy(data_skelet), volume_data=data)
# output = skan.filter_small_objects(data, 3)
skan.skeleton_analysis()
# self.assertEqual(output[5, 8, 7], 0)
skan.to_yaml(filename)
self.assertTrue(os.path.exists(filename))
