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 label_volumetric_vessel_tree(oseg,
vessel_label=None,
write_to_oseg=True,
new_label_str_format="{}{:03d}"):
"""
Split vessel by branches and put it in segmentation and slab.
:param oseg: OrganSegmentation object with segmentation, voxelsize_mm and slab
:param vessel_label: int or string label with vessel. Everything above zero is used if vessel_label is set None.
:param write_to_oseg: Store output into oseg.segmentation if True. The slab is also updated.
:param new_label_str_format: format of new slab
:return:
"""
import skelet3d
if vessel_label is None:
vessel_volume = oseg.segmentation > 0
else:
vessel_volume = oseg.select_label(vessel_label)
# print(np.unique(vessel_volume))
skel = skelet3d.skelet3d(vessel_volume)
skan = skelet3d.SkeletonAnalyser(skel, volume_data=vessel_volume)
skan.skeleton_analysis()
bl = skan.get_branch_label()
un = np.unique(bl)
if write_to_oseg:
for lb in un:
if lb != 0:
new_slabel = new_label_str_format.format(vessel_label, lb)
new_nlabel = oseg.nlabels(new_slabel)
oseg.segmentation[bl == lb] = new_nlabel
# ima.distance_segmentation(oseg.select_label(vessel_label))
return bl
def test_vessel_tree_vtk_from_skeleton_failing(self):
print("skelet3d_installed", skelet3d_installed)
import skelet3d
import skelet3d.skeleton_analyser
import shutil
fn_out = "tree.vtk"
if os.path.exists(fn_out):
os.remove(fn_out)
volume_data = np.zeros([3, 7, 9], dtype=np.int)
volume_data[:, :, 1:3] = 1
volume_data[:, 5, 2:9] = 1
volume_data[:, 0:7, 5] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(
skelet, volume_data=volume_data, voxelsize_mm=[1, 1, 1]
)
stats = skan.skeleton_analysis()
# tvg = TreeBuilder('vtk')
tvg = TBVTK()
tvg.set_model1d(stats)
tvg.voxelsize_mm = [1, 1, 1]
tvg.shape = [100, 100, 100]
tvg.tree_data = stats
output = tvg.buildTree() # noqa
tvg.saveToFile(fn_out)
os.path.exists(fn_out)
def test_skeleton_from_portal_vein(self):
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = io3d.datasets.generate_synthetic_liver()
skelet = skelet3d.skelet3d(segm == 2)
self.assertEqual(np.max(skelet), 1)
self.assertEqual(np.min(skelet), 0)
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_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_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))
def test_skelet3d(self):
data = np.zeros([8, 9, 10], dtype=np.int8)
data[1:4, 3:7, 1:12] = 1
# data[2:5, 2:8, 3:7] = 1
# data[2:5, 2:8, 3:5] = 1
skelet = skelet3d.skelet3d(data)
# expected result
expected_skelet = np.zeros([8, 9, 10], dtype=np.int8)
expected_skelet[2, 5, 2:9] = 1
self.assertTrue(np.array_equal(expected_skelet, skelet))
def binar_to_skeleton(self):
# create border with generated stuff for skeletonization
#expanded_data = self.create_border_for_skeletonization(self.data3d_thr, size=50)
expanded_data = self.data3d_thr
expanded_skel = skelet3d.skelet3d(
(expanded_data > 0).astype(np.int8)
)
# cut data shape back to original size
border = (expanded_data.shape[0]-self.data3d_thr.shape[0])/2
if border!=0:
self.data3d_skel = expanded_skel[border:-border, border:-border, border:-border].copy()
else:
self.data3d_skel = expanded_skel
def test_skeleton_problem_with_some_types_of_rectangular_structures(self):
# TODO fix bug in skeletonization algorithm.
# It ignores vertical structures and horizontal structures. Sometimes is just change the shape inough.
#
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = (
io3d.datasets.generate_synthetic_liver())
data_segm = segm == 2
# data_segm[39:44, 120:170, 100:111] = True
# data_segm[39:44, 50:120, 60:70] = True
# data_segm[39:44, 60:120, 100:105] = True
# data_segm[39:44, 60:120, 110:110 + 5] = True
data_segm[39:44, 120:130, 120:240] = True
data_segm[39:44, 60:120, 120:120 + 6] = True
data_segm[39:44, 60:120, 130:130 + 7] = True
data_segm[39:44, 60:120, 140:140 + 8] = True
data_segm[39:44, 60:120, 160:160 + 9] = True
data_segm[39:44, 60:120, 180:180 + 10] = True
data_segm[39:44, 60:120, 200:200 + 11] = True
data_segm[39:44, 60:120, 220:220 + 12] = True
# data_segm[39:44, 60:120, 120:126] = True
# data_segm[39:44, 60:120, 130:140] = True
# data_segm[39:44, 60:120, 150:151] = True
# data_segm[39:44, 60:120, 160:162] = True
data_skelet = skelet3d.skelet3d(data_segm)
import sed3
ed = sed3.sed3(data_skelet,
contour=data_segm) # , contour=branche_label)
ed.show()
self.assertEqual(
np.max(data_skelet[39:44, 60:100, 140:140 + 8]),
1,
"In this branche is expected skeleton",
)
self.assertEqual(np.min(data_skelet), 0)
self.assertEqual(np.max(data_skelet), 1)
def label_volumetric_vessel_tree(oseg,
vessel_label=None,
write_to_oseg=True,
new_label_str_format="{}{:03d}"):
"""
Split vessel by branches and put it in segmentation and slab.
:param oseg: OrganSegmentation object with segmentation, voxelsize_mm and slab
:param vessel_label: int or string label with vessel. Everything above zero is used if vessel_label is set None.
:param write_to_oseg: Store output into oseg.segmentation if True. The slab is also updated.
:param new_label_str_format: format of new slab
:return:
"""
logger.debug("vessel_label {}".format(vessel_label))
logger.debug("python version {} {}".format(sys.version_info,
sys.executable))
import skelet3d
if vessel_label is None:
vessel_volume = oseg.segmentation > 0
else:
vessel_volume = oseg.select_label(vessel_label)
# print(np.unique(vessel_volume))
skel = skelet3d.skelet3d(vessel_volume)
skan = skelet3d.SkeletonAnalyser(skel, volume_data=vessel_volume)
skan.skeleton_analysis()
bl = skan.get_branch_label()
un = np.unique(bl)
logger.debug("skelet3d branch label min: {}, max: {}, dtype: {}".format(
np.min(bl), np.max(bl), bl.dtype))
if write_to_oseg:
if 127 < np.max(bl) and ((oseg.segmentation.dtype == np.int8) or
(oseg.segmentation.dtype == np.uint8)):
oseg.segmentation = oseg.segmentation.astype(np.int16)
for lb in un:
if lb != 0:
new_slabel = new_label_str_format.format(vessel_label, lb)
new_nlabel = oseg.nlabels(new_slabel)
oseg.segmentation[bl == lb] = new_nlabel
# ima.distance_segmentation(oseg.select_label(vessel_label))
return bl
def __vesselTree(self, binaryData3d, textLabel):
import skelet3d
import skeleton_analyser # histology_analyser as skan
data3d_thr = (binaryData3d > 0).astype(np.int8)
data3d_skel = skelet3d.skelet3d(data3d_thr)
skan = skeleton_analyser.SkeletonAnalyser(
data3d_skel,
volume_data=data3d_thr,
voxelsize_mm=self.voxelsize_mm)
stats = skan.skeleton_analysis(guiUpdateFunction=None)
if 'graph' not in self.vessel_tree.keys():
self.vessel_tree['voxelsize_mm'] = self.voxelsize_mm
self.vessel_tree['graph'] = {}
self.vessel_tree['graph'][textLabel] = stats
# print sa.stats
# save skeleton to special file
misc.obj_to_file(self.vessel_tree, 'vessel_tree.yaml', filetype='yaml')
def test_vessel_tree_vtk_from_skeleton_with_more_tubes(self):
"""Test dont fail but there is no right output in vtk file"""
print("skelet3d_installed", skelet3d_installed)
import skelet3d
import skelet3d.skeleton_analyser
import shutil
fn_out = "tree_more_tubes.vtk"
if os.path.exists(fn_out):
os.remove(fn_out)
volume_data = np.zeros([20, 21, 22], dtype=np.int8)
# croess
volume_data[8:11, 14:17, 4:14] = 1
volume_data[9:15, 11:15, 9:19] = 1
volume_data[8:12, 5:9, 4:14] = 1
volume_data[9:15, 2:7, 9:19] = 1
# volume_data[10:12, 2:15, 13] = 1
# volume_data[11:13, 5:12, 14] = 1
# volume_data[12:14, 16, 13] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(
skelet,
volume_data=volume_data,
voxelsize_mm=[1, 1, 1],
cut_wrong_skeleton=False,
)
stats = skan.skeleton_analysis()
# self.assertEqual(len(stats), 1, "There should be just one cylinder based on data with different diameter")
# tvg = TreeBuilder('vtk')
tvg = TBVTK()
tvg.set_model1d(stats)
tvg.voxelsize_mm = [1, 1, 1]
tvg.shape = [100, 100, 100]
tvg.tree_data = stats
output = tvg.buildTree() # noqa
tvg.saveToFile(fn_out)
self.assertTrue(os.path.exists(fn_out))
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_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_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(data, 3)
# skan.skeleton_analysis()
# pe = ped.sed3(output)
# pe.show()
self.assertEqual(output[5, 8, 7], 0)
def test_donut(self):
print "donut"
# Create donut shape
data = np.ones([3, 7, 9])
data[:, 3, 3:6] = 0
expected_skelet = [
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]],
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 1, 1, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 1, 1, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]],
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]]]
skelet = skelet3d.skelet3d(data)
self.assertEqual(
0,
np.sum(np.abs(skelet - expected_skelet))
)
def test_skan_from_skeleton_of_one_tube(self):
import skelet3d.skeleton_analyser
# fn_out = 'tree_one_tube.vtk'
# if os.path.exists(fn_out):
# os.remove(fn_out)
volume_data = np.zeros([7, 8, 9], dtype=np.int)
volume_data[4:8, 4:6, 1:3] = 1
volume_data[:, 5, 2:9] = 1
volume_data[:, 0:7, 5] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(
skelet, volume_data=volume_data, voxelsize_mm=[1, 1, 1])
stats = skan.skeleton_analysis()
self.assertEqual(
len(stats),
1,
"There should be just one cylinder based on data with different diameter",
)
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_skeleton_problem_with_some_types_of_rectangular_structures(self):
# TODO fix bug in skeletonization algorithm.
# It ignores vertical structures and horizontal structures. Sometimes is just change the shape inough.
#
import io3d.datasets
data3d, segm, voxelsize_mm, slab, seeds_liver, seeds_porta = io3d.datasets.generate_synthetic_liver()
data_segm = segm == 2
# data_segm[39:44, 120:170, 100:111] = True
# data_segm[39:44, 50:120, 60:70] = True
# data_segm[39:44, 60:120, 100:105] = True
# data_segm[39:44, 60:120, 110:110 + 5] = True
data_segm[39:44, 120:130, 120:240] = True
data_segm[39:44, 60:120, 120:120 + 6] = True
data_segm[39:44, 60:120, 130:130 + 7] = True
data_segm[39:44, 60:120, 140:140 + 8] = True
data_segm[39:44, 60:120, 160:160 + 9] = True
data_segm[39:44, 60:120, 180:180 + 10] = True
data_segm[39:44, 60:120, 200:200 + 11] = True
data_segm[39:44, 60:120, 220:220 + 12] = True
# data_segm[39:44, 60:120, 120:126] = True
# data_segm[39:44, 60:120, 130:140] = True
# data_segm[39:44, 60:120, 150:151] = True
# data_segm[39:44, 60:120, 160:162] = True
data_skelet = skelet3d.skelet3d(data_segm)
import sed3
ed = sed3.sed3(data_skelet, contour=data_segm)# , contour=branche_label)
ed.show()
self.assertEqual(np.max(
data_skelet[39:44, 60:100, 140:140 + 8]
), 1, "In this branche is expected skeleton")
self.assertEqual(np.min(data_skelet), 0)
self.assertEqual(np.max(data_skelet), 1)
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))
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_donut(self):
# Create donut shape
data = np.ones([3, 7, 9])
data[:, 3, 3:6] = 0
expected_skelet = [
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]],
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 1, 1, 1, 1, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 1, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 1, 1, 1, 1, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]],
[
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 0]]]
skelet = skelet3d.skelet3d(data)
self.assertEqual(
0,
np.sum(np.abs(skelet - expected_skelet))
)
def label_volumetric_vessel_tree(oseg, vessel_label=None, write_to_oseg=True, new_label_str_format="{}{:03d}"):
"""
Split vessel by branches and put it in segmentation and slab.
:param oseg: OrganSegmentation object with segmentation, voxelsize_mm and slab
:param vessel_label: int or string label with vessel. Everything above zero is used if vessel_label is set None.
:param write_to_oseg: Store output into oseg.segmentation if True. The slab is also updated.
:param new_label_str_format: format of new slab
:return:
"""
logger.debug("vessel_label {}".format(vessel_label))
logger.debug("python version {} {}".format(sys.version_info, sys.executable))
import skelet3d
if vessel_label is None:
vessel_volume = oseg.segmentation > 0
else:
vessel_volume = oseg.select_label(vessel_label)
# print(np.unique(vessel_volume))
skel = skelet3d.skelet3d(vessel_volume)
skan = skelet3d.SkeletonAnalyser(skel, volume_data=vessel_volume)
skan.skeleton_analysis()
bl = skan.get_branch_label()
un = np.unique(bl)
logger.debug("skelet3d branch label min: {}, max: {}, dtype: {}".format(np.min(bl), np.max(bl), bl.dtype))
if write_to_oseg:
if 127 < np.max(bl) and ((oseg.segmentation.dtype == np.int8) or (oseg.segmentation.dtype == np.uint8)):
oseg.segmentation = oseg.segmentation.astype(np.int16)
for lb in un:
if lb != 0:
new_slabel = new_label_str_format.format(vessel_label, lb)
new_nlabel = oseg.nlabels(new_slabel)
oseg.segmentation[bl == lb] = new_nlabel
# ima.distance_segmentation(oseg.select_label(vessel_label))
return bl
def test_vessel_tree_vtk_from_skeleton_of_one_tube(self):
print("skelet3d_installed", skelet3d_installed)
import skelet3d
import skelet3d.skeleton_analyser
import shutil
fn_out = "tree_one_tube.vtk"
if os.path.exists(fn_out):
os.remove(fn_out)
volume_data = np.zeros([7, 8, 9], dtype=np.int)
volume_data[4:8, 4:6, 1:3] = 1
volume_data[:, 5, 2:9] = 1
volume_data[:, 0:7, 5] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(
skelet, volume_data=volume_data, voxelsize_mm=[1, 1, 1]
)
stats = skan.skeleton_analysis()
self.assertEqual(
len(stats),
1,
"There should be just one cylinder based on data with different diameter",
)
# tvg = TreeBuilder('vtk')
tvg = TBVTK()
tvg.set_model1d(stats)
tvg.voxelsize_mm = [1, 1, 1]
tvg.shape = [100, 100, 100]
tvg.tree_data = stats
output = tvg.buildTree() # noqa
tvg.saveToFile(fn_out)
self.assertTrue(os.path.exists(fn_out))
#! /usr/bin/env python # -*- coding: utf-8 -*- # vim:fenc=utf-8 # # Copyright © %YEAR% %USER% <%MAIL%> # # Distributed under terms of the %LICENSE% license. """ %HERE% """ import skelet3d import numpy as np # Create donut shape data = np.ones([3, 7, 9]) data[:, 3, 3:6] = 0 skelet = skelet3d.skelet3d(data) print skelet
#! /usr/bin/env python
# -*- coding: utf-8 -*-
# vim:fenc=utf-8
import numpy as np
import skelet3d
import skelet3d.skeleton_analyser
import skelet3d.tree
from skelet3d.tree import TreeBuilder
# output vtk file can be visualized with ParaView
fn_out = 'tree.vkt'
volume_data = np.zeros([3, 7, 9], dtype=np.int)
volume_data [:, :, 1:3] = 1
volume_data [:, 5, 2:9] = 1
volume_data [:, 0:7, 5] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(skelet, volume_data=volume_data, voxelsize_mm=[1,1,1])
stats = skan.skeleton_analysis()
tvg = TreeBuilder('vtk')
tvg.voxelsize_mm = [1, 1, 1]
tvg.shape = [100, 100, 100]
tvg.tree_data = stats
output = tvg.buildTree() # noqa
tvg.saveToFile(fn_out)
print("Output saved as '{}'".format(fn_out))
# -*- coding: utf-8 -*-
# vim:fenc=utf-8
import numpy as np
import skelet3d
import skelet3d.skeleton_analyser
import skelet3d.tree
from skelet3d.tree import TreeBuilder
# output vtk file can be visualized with ParaView
fn_out = 'tree.vkt'
volume_data = np.zeros([3, 7, 9], dtype=np.int)
volume_data[:, :, 1:3] = 1
volume_data[:, 5, 2:9] = 1
volume_data[:, 0:7, 5] = 1
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.skeleton_analyser.SkeletonAnalyser(skelet,
volume_data=volume_data,
voxelsize_mm=[1, 1, 1])
stats = skan.skeleton_analysis()
tvg = TreeBuilder('vtk')
tvg.voxelsize_mm = [1, 1, 1]
tvg.shape = [100, 100, 100]
tvg.tree_data = stats
output = tvg.buildTree() # noqa
tvg.saveToFile(fn_out)
print("Output saved as '{}'".format(fn_out))
def voda_sk(organ_seg, liver_seg, voxelsize, cr):
"""
build skeleton of porta
voda_sk(xyz) build distance 3D map and compute volumes of liverpart for each skeleton element of chosen dataset
voda_sk(xyz) takes argument from load_vdata - it need volumetric data of porta, liver and vox size (not ncsry..)
"""
l_d = [organ_seg, liver_seg, voxelsize, cr]
tempsz = (np.shape(l_d[0][1]))
ylab = tempsz[1]
xlab = tempsz[0]
volume_data = l_d[0]
#skeletonization - params are saved in var stats
skelet = skelet3d.skelet3d(volume_data)
skan = skelet3d.SkeletonAnalyser(skelet, volume_data = volume_data, voxelsize_mm = [1, 1, 1])
stats = skan.skeleton_analysis()
edge_number = 1
#build of 3D distance map
linesarr_x = []
linesarr_y = []
linesarr_z = []
temparr = skan.sklabel
counter = 0
tempslx_l = 0
for i in range(0,len(l_d[0])):
for j in range(0, xlab):
for k in range(0, ylab):
if (temparr[i][j][k] != 0):
linesarr_x.append(k)
linesarr_y.append(j)
counter = counter + 1
for z in range(0, counter):
tempslx_l = i * l_d[2] #third value in l_data is space between slices of current datasete
linesarr_z.append(tempslx_l)
counter = 0
distance_map = imlb.distance_segmentation(temparr)
nodeArr = []
for i in range(1, (len(stats)) + 1):
#print(i)
try:
nodeArr.append((stats[i]['nodeA_ZYX']))
except:
None
try:
nodeArr.append((stats[i]['nodeB_ZYX']))
except:
None
zdata = []
xdata = []
ydata = []
temp_xp = 2
temp_yp = 1
temp_zp = 0
for i in range(len(nodeArr)):
zdata.append(nodeArr[i][temp_zp])
ydata.append(nodeArr[i][temp_yp])
xdata.append(nodeArr[i][temp_xp])
zdata[i] = zdata[i]*l_d[2]
#build list of all elements in skeleton
list_= []
for slx in range(len(l_d[1])):
for x in range(0, xlab):
for y in range(0, ylab):
if(distance_map[slx, x, y]) < 1e5:
list_.append(distance_map[slx, x, y])
set(list_)
list_of_areas = set(list_)
list_of_areas_arr = list(list_of_areas)
list_of_areas_arr_edges = []
for x in range(0, len(list_of_areas_arr)):
if(list_of_areas_arr[x] > 0):
list_of_areas_arr_edges.append(list_of_areas_arr[x])
#plot data if needed
slicearr_p = []
zte_p = []
xte_p = []
yte_p = []
counter = 0
for slx in range(len(l_d[0])):
for x in range(0, xlab):
for y in range(0, ylab):
if(l_d[0][slx, x, y]) == 1:
xte_p.append(y)
yte_p.append(x)
counter = counter + 1
for z in range(0, counter):
tempslx_p = slx * l_d[2]
zte_p.append(tempslx_p)
counter = 0
slicearr_p.append(xte_p)
slicearr_p.append(yte_p)
slicearr_p.append(zte_p)
slicearr_l = []
zte_l = []
xte_l = []
yte_l = []
counter = 0
for slx in range(len(l_d[1])):
for x in range(0, xlab):
for y in range(0, ylab):
if(l_d[1][slx, x, y]) == 1:
xte_l.append(y)
yte_l.append(x)
counter = counter + 1
for z in range(0, counter):
tempslx_l = slx * l_d[2]
zte_l.append(tempslx_l)
counter = 0
slicearr_l.append(xte_l)
slicearr_l.append(yte_l)
slicearr_l.append(zte_l)
dist_map_x = []
dist_map_y = []
dist_map_z = []
counter_sl = 0
#compute of ALL VOLUMES - takes time - control print for each skelet element
dist_map_final_liver_vol_temp = []
dist_map_final_liver_vol = []
loar_l = len(list_of_areas_arr)
if(loar_l % 2 == 0):
loar_l = loar_l/2
else:
loar_ = loar_l/2 + 1
for area in range(0, loar_l):
temp_area = list_of_areas_arr[area]
for slc in range(0, len(l_d[1])):
for x in range(0, xlab):
for y in range(0, ylab):
if ((l_d[1][slc, x, y]) == 1):
if(((distance_map[slc, x, y]) == temp_area)):
dist_map_x.append(y)
dist_map_y.append(x)
counter_sl = counter_sl + 1
for slic in range(0, counter_sl):
dist_map_z.append(slc)
counter_sl = 0
dist_map_final_liver_vol_temp.append(dist_map_z)
dist_map_final_liver_vol_temp.append(dist_map_y)
dist_map_final_liver_vol_temp.append(dist_map_x)
dist_map_final_liver_vol_temp.append(temp_area)
dist_map_final_liver_vol.append(dist_map_final_liver_vol_temp)
dist_map_final_liver_vol_temp = []
dist_map_x = []
dist_map_y = []
dist_map_z = []
print("computing volume of area: ", temp_area, "Num. of areas in this dataset: ", len(list_of_areas_arr))
print("$$$$$-------------------------------------------------------------------$$$$$")
# ret_array = []
# ret_array.append(stats)
# ret_array.append(list_of_areas_arr_edges)
# ret_array.append(dist_map_final_liver_vol)
return stats, list_of_areas_arr_edges, dist_map_final_liver_vol
import os
import os.path
import pytest
path_to_script = os.path.dirname(os.path.abspath(__file__))
import unittest
import numpy as np
import sys
try:
import skelet3d
data3d = np.ones([3, 7, 9])
data3d[:, 3, 3:6] = 0
skelet3d.skelet3d(data3d)
skelet3d_installed = True
# skelet3d
except:
skelet3d_installed = False
logger.warning("skelet3d is not working")
try:
import larcc
larcc_installed = True
except:
larcc_installed = False
logger.warning("larcc is not working")
from fibrous.tree import TreeBuilder
#! /usr/bin/env python # -*- coding: utf-8 -*- # vim:fenc=utf-8 # # Copyright © %YEAR% %USER% <%MAIL%> # # Distributed under terms of the %LICENSE% license. """ %HERE% """ import skelet3d import numpy as np # Create donut shape data = np.ones([3,7,9]) data [:, 3, 3:6] = 0 skelet = skelet3d.skelet3d(data) print(skelet)
def binar_to_skeleton(self, data3d_thr):
data3d_thr = (data3d_thr > 0).astype(np.int8)
data3d_skel = skelet3d.skelet3d(data3d_thr)
return data3d_skel
