def test_remove_small_regions():
data = np.array([[[0., 1., 0.], [0., 1., 1.], [0., 0., 0.]],
[[0., 0., 0.], [1., 0., 0.], [0., 1., 0.]],
[[0., 0., 1.], [1., 0., 0.], [0., 1., 1.]]])
# To remove small regions, data should be labelled
label_map, n_labels = ndimage.label(data)
sum_label_data = np.sum(label_map)
affine = np.eye(4)
min_size = 10
# data can be act as mask_data to identify regions in label_map because
# features in label_map are built upon non-zeros in data
index = np.arange(n_labels + 1)
removed_data = _remove_small_regions(label_map, index, affine, min_size)
sum_removed_data = np.sum(removed_data)
assert sum_removed_data < sum_label_data
def test_remove_small_regions():
data = np.array([[[0., 1., 0.],
[0., 1., 1.],
[0., 0., 0.]],
[[0., 0., 0.],
[1., 0., 0.],
[0., 1., 0.]],
[[0., 0., 1.],
[1., 0., 0.],
[0., 1., 1.]]])
# To remove small regions, data should be labelled
label_map, n_labels = ndimage.label(data)
sum_label_data = np.sum(label_map)
affine = np.eye(4)
min_size = 10
# data can be act as mask_data to identify regions in label_map because
# features in label_map are built upon non-zeros in data
index = np.arange(n_labels + 1)
removed_data = _remove_small_regions(label_map, index, affine, min_size)
sum_removed_data = np.sum(removed_data)
assert_true(sum_removed_data < sum_label_data)