def __run_boundary_difference_of_means_test(self,
label,
gradient,
expected_result,
msg=''):
label = numpy.asarray(label)
gradient = numpy.asarray(gradient)
graph = GCGraphTest(
numpy.unique(label).size, math.pow(numpy.unique(label).size, 2))
boundary_difference_of_means(graph, label, (gradient))
graph.validate_nweights(self, expected_result, msg)
def test_boundary_difference_of_means_3d(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() function for 3D."""
# 3D VERSION
# the gradient magnitude image
original = [[[0., 0., 0., 0.1, 0.1, 0.5], [0., 0., 0., 0.1, 0.1, 0.1],
[0., 0.3, 0.3, 0.2, 0.2,
0.2], [0., 0., 0.3, 0.3, 0.3, 0.2],
[0., 0., 0.3, 0.2, 0.2, 0.4]],
[[0., 0., 0., 0.1, 0.1, 0.6], [0., 0., 0., 0.1, 0.1, 0.1],
[0., 0.3, 0.3, 0.2, 0.2,
0.2], [0., 0., 0.3, 0.3, 0.3, 0.2],
[0., 0., 0.3, 0.2, 0.2, 0.4]]]
# the label image (labels have to start from 0)
label = [[[0, 0, 0, 1, 1, 5], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2],
[0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]],
[[0, 0, 0, 1, 1, 6], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2],
[0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]]]
# the expected result
expected_result = {
(0, 1): 0.833333333,
(0, 3): 0.5,
(1, 2): 0.833333333,
(1, 5): 0.333333333,
(1, 6): 0.166666667,
(2, 3): 0.833333333,
(2, 4): 0.666666667,
(5, 6): 0.833333333
} # only 3D edge
expected_result = self.__to_two_directed(expected_result)
# run the function
result = boundary_difference_of_means(label, (original))
result = self._reorder_keys(result)
# check returned values
self.__compare_dictionaries(result, expected_result)
def test_boundary_difference_of_means_3d(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() function for 3D."""
# 3D VERSION
# the gradient magnitude image
original = [
[
[0.0, 0.0, 0.0, 0.1, 0.1, 0.5],
[0.0, 0.0, 0.0, 0.1, 0.1, 0.1],
[0.0, 0.3, 0.3, 0.2, 0.2, 0.2],
[0.0, 0.0, 0.3, 0.3, 0.3, 0.2],
[0.0, 0.0, 0.3, 0.2, 0.2, 0.4],
],
[
[0.0, 0.0, 0.0, 0.1, 0.1, 0.6],
[0.0, 0.0, 0.0, 0.1, 0.1, 0.1],
[0.0, 0.3, 0.3, 0.2, 0.2, 0.2],
[0.0, 0.0, 0.3, 0.3, 0.3, 0.2],
[0.0, 0.0, 0.3, 0.2, 0.2, 0.4],
],
]
# the label image (labels have to start from 0)
label = [
[[0, 0, 0, 1, 1, 5], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2], [0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]],
[[0, 0, 0, 1, 1, 6], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2], [0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]],
]
# the expected result
expected_result = {
(0, 1): 0.833333333,
(0, 3): 0.5,
(1, 2): 0.833333333,
(1, 5): 0.333333333,
(1, 6): 0.166666667,
(2, 3): 0.833333333,
(2, 4): 0.666666667,
(5, 6): 0.833333333,
} # only 3D edge
expected_result = self.__to_two_directed(expected_result)
# run the function
result = boundary_difference_of_means(label, (original))
result = self._reorder_keys(result)
# check returned values
self.__compare_dictionaries(result, expected_result)
def test_boundary_difference_of_means_2d(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() function for 2D."""
# the original image
original = [
[0.0, 0.0, 0.0, 0.1, 0.1, 0.5],
[0.0, 0.0, 0.0, 0.1, 0.1, 0.1],
[0.0, 0.3, 0.3, 0.2, 0.2, 0.2],
[0.0, 0.0, 0.3, 0.3, 0.3, 0.2],
[0.0, 0.0, 0.3, 0.2, 0.2, 0.4],
]
# the label image (labels have to start from 0)
label = [[0, 0, 0, 1, 1, 5], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2], [0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]]
# the expected result
expected_result = {(0, 1): 0.8, (0, 3): 0.4, (1, 2): 0.8, (1, 5): 0.2, (2, 3): 0.8, (2, 4): 0.6}
expected_result = self.__to_two_directed(expected_result)
# run the function
result = boundary_difference_of_means(label, (original))
result = self._reorder_keys(result)
# check returned values
self.__compare_dictionaries(result, expected_result)
def test_boundary_difference_of_means_2d(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() function for 2D."""
# the original image
original = [[0., 0., 0., 0.1, 0.1, 0.5], [0., 0., 0., 0.1, 0.1, 0.1],
[0., 0.3, 0.3, 0.2, 0.2, 0.2],
[0., 0., 0.3, 0.3, 0.3, 0.2], [0., 0., 0.3, 0.2, 0.2, 0.4]]
# the label image (labels have to start from 0)
label = [[0, 0, 0, 1, 1, 5], [0, 0, 0, 1, 1, 1], [0, 3, 3, 2, 2, 2],
[0, 0, 3, 3, 3, 2], [0, 0, 3, 2, 2, 4]]
# the expected result
expected_result = {
(0, 1): 0.8,
(0, 3): 0.4,
(1, 2): 0.8,
(1, 5): 0.2,
(2, 3): 0.8,
(2, 4): 0.6
}
expected_result = self.__to_two_directed(expected_result)
# run the function
result = boundary_difference_of_means(label, (original))
result = self._reorder_keys(result)
# check returned values
self.__compare_dictionaries(result, expected_result)
def test_boundary_difference_of_means_borders(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() border conditions."""
# TEST1: test for a label image with not continuous label ids not starting from 0
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test1')
# TEST2: test for a label image with negative labels
label = [[-1, 4, 8], [-1, 3, 10], [1, -3, 10]]
expected_result = {
(-1, 1): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(-1, 3): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(-3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(-3, 1): (sys.float_info.min, sys.float_info.min),
(-3, 3): (sys.float_info.min, sys.float_info.min),
(-1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test2')
# TEST3: test for behavior on occurrence of very small (~0) and 1 weights
gradient = [[0., 0., 0.], [0., 0., sys.float_info.max]]
label = [[0, 1, 2], [0, 1, 3]]
expected_result = {
(0, 1): (1.0, 1.0),
(1, 2): (1.0, 1.0),
(1, 3): (sys.float_info.min, sys.float_info.min),
(2, 3): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(label, (gradient))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test3')
# TEST4: check behavior for integer gradient image
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
label = scipy.asarray(label)
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(
label, (scipy.zeros(label.shape, scipy.int_)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test4')
# TEST5: reaction to different array orders
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
label = scipy.asarray(label, order='C') # C-order, gradient same order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test5 (C,C)')
label = scipy.asarray(label,
order='F') # Fortran order, gradient same order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test5 (F, F)')
label = scipy.asarray(label,
order='C') # C-order, gradient different order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(
label, (scipy.zeros(label.shape, order='F')))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test5 (C, F)')
label = scipy.asarray(label,
order='F') # F-order, gradient different order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min)
}
result = boundary_difference_of_means(
label, (scipy.zeros(label.shape, order='C')))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, 'Test5 (F, C)')
def test_boundary_difference_of_means_borders(self):
"""Test the @link medpy.graphcut.boundary_difference_of_means() border conditions."""
# TEST1: test for a label image with not continuous label ids not starting from 0
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test1")
# TEST2: test for a label image with negative labels
label = [[-1, 4, 8], [-1, 3, 10], [1, -3, 10]]
expected_result = {
(-1, 1): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(-1, 3): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(-3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(-3, 1): (sys.float_info.min, sys.float_info.min),
(-3, 3): (sys.float_info.min, sys.float_info.min),
(-1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test2")
# TEST3: test for behavior on occurrence of very small (~0) and 1 weights
gradient = [[0.0, 0.0, 0.0], [0.0, 0.0, sys.float_info.max]]
label = [[0, 1, 2], [0, 1, 3]]
expected_result = {
(0, 1): (1.0, 1.0),
(1, 2): (1.0, 1.0),
(1, 3): (sys.float_info.min, sys.float_info.min),
(2, 3): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (gradient))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test3")
# TEST4: check behavior for integer gradient image
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
label = scipy.asarray(label)
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros(label.shape, scipy.int_)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test4")
# TEST5: reaction to different array orders
label = [[1, 4, 8], [1, 3, 10], [1, 3, 10]]
label = scipy.asarray(label, order="C") # C-order, gradient same order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test5 (C,C)")
label = scipy.asarray(label, order="F") # Fortran order, gradient same order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros_like(label)))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test5 (F, F)")
label = scipy.asarray(label, order="C") # C-order, gradient different order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros(label.shape, order="F")))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test5 (C, F)")
label = scipy.asarray(label, order="F") # F-order, gradient different order
expected_result = {
(1, 3): (sys.float_info.min, sys.float_info.min),
(4, 8): (sys.float_info.min, sys.float_info.min),
(3, 10): (sys.float_info.min, sys.float_info.min),
(8, 10): (sys.float_info.min, sys.float_info.min),
(1, 4): (sys.float_info.min, sys.float_info.min),
(3, 4): (sys.float_info.min, sys.float_info.min),
}
result = boundary_difference_of_means(label, (scipy.zeros(label.shape, order="C")))
result = self._reorder_keys(result)
self.__compare_dictionaries(result, expected_result, "Test5 (F, C)")
