def _make_features(self,
features: str,
labels,
annotation=None,
image=None):
import pyclesperanto_prototype as cle
pixel_statistics = cle.statistics_of_background_and_labelled_pixels(
image, labels)
if annotation is not None:
# determine ground truth
annotation_statistics = cle.statistics_of_background_and_labelled_pixels(
annotation, labels)
classification_gt = annotation_statistics['max_intensity']
classification_gt[0] = 0
else:
classification_gt = None
feature_list = features.split(',')
table, gt = self._select_features(pixel_statistics, feature_list,
labels, classification_gt)
return table, gt
def test_push_regionprops_column():
labels = cle.push(np.asarray([[
[4, 4, 2],
], [
[4, 1, 2],
], [[3, 3, 4]]]))
reference = [0, 1, 2, 2, 4]
regionprops = cle.statistics_of_background_and_labelled_pixels(
None, labels)
areas = cle.push_regionprops_column(regionprops, 'area')
print(reference)
print(areas)
assert np.allclose(reference, areas, 0.001)
def test_statistics_of_background_and_labelled_pixels():
intensity = cle.push(np.asarray([[0, 1, 2], [1, 2, 3], [2, 3, 4]]))
labels = cle.push(np.asarray([[0, 0, 1], [0, 1, 1], [2, 2, 2]]))
reference = cle.push(
np.asarray([
#IDENTIFIER(0),
[0., 1., 2.],
# BOUNDING_BOX_X(1),
[0., 1., 0.],
# BOUNDING_BOX_Y(2),
[0., 0., 2.],
# BOUNDING_BOX_Z(3),
[0., 0., 0.],
# BOUNDING_BOX_END_X(4),
[1., 2., 2.],
# BOUNDING_BOX_END_Y(5),
[1., 1., 2.],
# BOUNDING_BOX_END_Z(6),
[0., 0., 0.],
# BOUNDING_BOX_WIDTH(7),
[2., 2., 3.],
# BOUNDING_BOX_HEIGHT(8),
[2., 2., 1.],
# BOUNDING_BOX_DEPTH(9),
[1., 1., 1.],
# MINIMUM_INTENSITY(10),
[0., 2., 2.],
# MAXIMUM_INTENSITY(11),
[1., 3., 4.],
# MEAN_INTENSITY(12)
[0.6666667, 2.3333333, 3.],
# SUM_INTENSITY(13),
[2., 7., 9.],
# STANDARD_DEVIATION_INTENSITY(14),
[0.47140452, 0.4714045, 0.8164966],
# PIXEL_COUNT(15),
[3., 3., 3.],
# SUM_INTENSITY_TIMES_X(16),
[1., 12., 11.],
# SUM_INTENSITY_TIMES_Y(17),
[1., 5., 18.],
# SUM_INTENSITY_TIMES_Y(17),
[0., 0., 0.],
# MASS_CENTER_X(19),
[0.5, 1.7142857, 1.2222222],
# MASS_CENTER_Y(20),
[0.5, 0.71428573, 2.],
# MASS_CENTER_Z(21),
[0., 0., 0.],
# SUM_X(22),
[1., 5., 3.],
# SUM_Y(23),
[1., 2., 6.],
# SUM_Z(24),
[0., 0., 0.],
# CENTROID_X(25),
[0.33333334, 1.6666666, 1.],
# CENTROID_Y(26),
[0.33333334, 0.6666667, 2.],
# CENTROID_Z(27),
[0., 0., 0.],
# SUM_DISTANCE_TO_MASS_CENTER(28),
[2.1213202, 1.9426796, 2.2222223],
# MEAN_DISTANCE_TO_MASS_CENTER(29),
[0.70710677, 0.6475599, 0.7407408],
# MAX_DISTANCE_TO_MASS_CENTER(30),
[0.70710677, 0.7693094, 1.2222222],
# MAX_MEAN_DISTANCE_TO_MASS_CENTER_RATIO(31),
[1., 1.1880127, 1.6499999],
# SUM_DISTANCE_TO_CENTROID(32),
[1.9621165, 1.9621165, 2.],
# MEAN_DISTANCE_TO_CENTROID(33),
[0.65403885, 0.65403885, 0.6666667],
# MAX_DISTANCE_TO_CENTROID(34),
[0.74535596, 0.745356, 1.],
# MAX_MEAN_DISTANCE_TO_CENTROID_RATIO(35);
[1.1396203, 1.1396204, 1.5]
]))
result = cle.statistics_of_background_and_labelled_pixels(
intensity, labels)
result_image = cle.push_regionprops(result, first_row_index=0)
a = cle.pull(result_image)
b = cle.pull(reference)
print(a)
print(b)
assert (np.allclose(a, b, 0.0001))
def test_statistics_of_background_and_labelled_pixels_gpu_compare_to_clij2():
from skimage.io import imread
blobs = cle.push(imread('data/mini_blobs.tif'))
labels = cle.push(
imread('data/mini_blobs_otsu_labels_excluded_edges_imagej.tif'))
regionprops = cle.statistics_of_background_and_labelled_pixels(
blobs, labels)
table = cle.pull(cle.transpose_xy(cle.push_regionprops(regionprops)))
print(table)
import numpy as np
# np.savetxt('../data/mini_blobs_measurements_pyclesperanto.csv', table,delimiter=',')
clij_reference_table = np.loadtxt('data/mini_blobs_measurements_clij.csv',
delimiter=',',
skiprows=1)
# chop off first column as ImageJ saved an additional counter column
clij_reference_table = clij_reference_table[:, 1:]
print(clij_reference_table)
columns_to_test = [
cle.STATISTICS_ENTRY.IDENTIFIER.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_X.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_Y.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_Z.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_END_X.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_END_Y.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_END_Z.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_WIDTH.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_HEIGHT.value,
cle.STATISTICS_ENTRY.BOUNDING_BOX_DEPTH.value,
cle.STATISTICS_ENTRY.MINIMUM_INTENSITY.value,
cle.STATISTICS_ENTRY.MAXIMUM_INTENSITY.value,
cle.STATISTICS_ENTRY.MEAN_INTENSITY.value,
cle.STATISTICS_ENTRY.SUM_INTENSITY.value,
cle.STATISTICS_ENTRY.STANDARD_DEVIATION_INTENSITY.value,
cle.STATISTICS_ENTRY.PIXEL_COUNT.value,
cle.STATISTICS_ENTRY.MASS_CENTER_X.value,
cle.STATISTICS_ENTRY.MASS_CENTER_Y.value,
cle.STATISTICS_ENTRY.MASS_CENTER_Z.value,
cle.STATISTICS_ENTRY.CENTROID_X.value,
cle.STATISTICS_ENTRY.CENTROID_Y.value,
cle.STATISTICS_ENTRY.CENTROID_Z.value,
cle.STATISTICS_ENTRY.SUM_DISTANCE_TO_CENTROID.value,
cle.STATISTICS_ENTRY.MEAN_DISTANCE_TO_CENTROID.value,
cle.STATISTICS_ENTRY.MAX_DISTANCE_TO_CENTROID.value,
cle.STATISTICS_ENTRY.MAX_MEAN_DISTANCE_TO_CENTROID_RATIO.value,
cle.STATISTICS_ENTRY.SUM_INTENSITY_TIMES_X.value,
cle.STATISTICS_ENTRY.SUM_INTENSITY_TIMES_Y.value,
cle.STATISTICS_ENTRY.SUM_INTENSITY_TIMES_Z.value,
cle.STATISTICS_ENTRY.SUM_X.value,
cle.STATISTICS_ENTRY.SUM_Y.value,
cle.STATISTICS_ENTRY.SUM_Z.value,
cle.STATISTICS_ENTRY.SUM_DISTANCE_TO_MASS_CENTER.value,
cle.STATISTICS_ENTRY.MEAN_DISTANCE_TO_MASS_CENTER.value,
cle.STATISTICS_ENTRY.MAX_DISTANCE_TO_MASS_CENTER.value,
cle.STATISTICS_ENTRY.MAX_MEAN_DISTANCE_TO_MASS_CENTER_RATIO.value,
]
for column_index in columns_to_test:
reference_values = clij_reference_table[:, column_index]
values = table[:, column_index]
print("testing STATISTICS_ENTRY" +
str(cle.STATISTICS_ENTRY(column_index)))
print(reference_values)
print(values)
assert np.allclose(reference_values, values, 0.001)
