def test_wrong_dtype():
with pytest.raises(ValueError):
transform(
np.ones((1, )),
np.eye(1),
(0, ),
output_image=np.zeros((1, ), dtype=np.float32),
)
def test_wrong_origin_dimension():
for dim in range(1, 6):
image = np.ones((1, ) * dim)
for d in (x for x in range(1, 6) if x != dim):
with pytest.raises(ValueError):
transform(image,
np.eye(dim),
translation=(0, ) * dim,
origin=(0, ) * d)
def test_wrong_output_image_dim():
for dim in range(1, 6):
image = np.ones((1, ) * dim)
for d in (x for x in range(1, 6) if x != dim):
output = np.ones((1, ) * d)
with pytest.raises(ValueError):
transform(image,
np.eye(dim),
translation=(0, ) * dim,
output_image=output)
def test_float32():
image = np.ones((1, ), dtype=np.float32)
output_image = np.zeros((1, ), dtype=np.float32)
output = transform(image, np.eye(1), (0, ), output_image=output_image)
assert output == 1
assert output.dtype == np.float32
assert output_image == 1
def test_data_type_conversion():
image = np.ones((1, ), dtype=int)
output = transform(image,
np.eye(1),
translation=(0, ),
output_image_origin=(0.5, ),
order="linear")
assert output == 0.5
def test_translation_by_third():
for dim in range(1, 6):
image = np.ones((6, ) * dim) * 2
for select_dim in range(dim):
translation = [0] * dim
translation[select_dim] = 2 / 3
indices = [slice(None, None)] * dim
indices[select_dim] = slice(1)
expected_output = np.ones((6, ) * dim) * 2
expected_output[tuple(indices)] = 1 / 3 * 2
output = transform(image, np.eye(dim), translation=translation)
np.testing.assert_allclose(output, expected_output)
def test_translation_by_whole_number():
for dim in range(1, 6):
image = np.ones((6, ) * dim)
for select_dim in range(dim):
translation = [0] * dim
translation[select_dim] = 2
indices = [slice(None, None)] * dim
indices[select_dim] = slice(2)
expected_output = np.ones((6, ) * dim)
expected_output[tuple(indices)] = 0
output = transform(image, np.eye(dim), translation=translation)
np.testing.assert_allclose(output, expected_output)
def test_extract_slice():
image = np.zeros((6, ) * 3)
image[(slice(3, 6), ) * 3] = 1
output_slice = np.empty((6, 6, 1))
output = transform(
image,
np.eye(3),
(0, 0, 0),
output_image=output_slice,
output_image_origin=(0, 0, 3),
)
np.testing.assert_allclose(output[:, :, 0], image[:, :, 3])
def test_background_value():
for dim in range(1, 6):
image = np.ones((6, ) * dim) * 2
for select_dim in range(dim):
translation = [0] * dim
translation[select_dim] = 2 / 3
indices = [slice(None, None)] * dim
indices[select_dim] = slice(1)
expected_output = np.ones((6, ) * dim) * 2
expected_output[tuple(indices)] = 1 / 3 * 2 + 2 / 3 * 15
output = transform(image,
np.eye(dim),
translation=translation,
background_value=15.0)
np.testing.assert_allclose(output, expected_output)
def test_wrong_dimension_linear_transform():
for dim in range(1, 6):
image = np.ones((1, ) * dim)
for d in (x for x in range(1, 6) if x != dim):
matrix = np.eye(5)
with pytest.raises(ValueError):
transform(image, matrix[:dim, :d], translation=(0, ) * dim)
with pytest.raises(ValueError):
transform(image, matrix[:d, :dim], translation=(0, ) * dim)
with pytest.raises(ValueError):
transform(image, matrix[:d, :d], translation=(0, ) * dim)
def test_cubic_interpolation():
image = np.zeros((9, ) * 1)
image[3] = 13.3
image_test = transform(image, np.eye(1), (1.7, ), order="cubic")
def catmull_rom_interp(p0, p1, p2, p3, x):
return ((-0.5 * p0 + 1.5 * p1 - 1.5 * p2 + 0.5 * p3) * (x**3) +
(p0 - 2.5 * p1 + 2 * p2 - 0.5 * p3) * (x**2) +
(-0.5 * p0 + 0.5 * p2) * x + p1)
image_padded = np.pad(image, 3, mode="constant", constant_values=0)
result_manual = np.zeros((9, ))
for i in range(9):
result_manual[i] = catmull_rom_interp(*image_padded[i:i + 4], 0.3)
np.testing.assert_allclose(image_test, result_manual, atol=1e-10)
def test_rotation_multiple_90():
for dim in range(2, 6):
image_1 = np.ones((6, ) * dim)
image_2 = np.zeros((6, ) * dim)
image_2[(slice(None, 3), ) * dim] = 1
center_1 = None
center_2 = (1, ) * dim
for image, center in zip((image_1, image_2), (center_1, center_2)):
v1 = np.zeros(dim)
v2 = np.zeros(dim)
v1[-2] = 1
v2[-1] = 1
for angle in (np.pi / 2 * x for x in range(1, 4)):
rotation = mgen.rotation_from_angle_and_plane(angle, v1, v2)
output = transform(image, rotation, (0, ) * dim, origin=center)
# need some atol here since we are comparing exactly to 0
np.testing.assert_allclose(output, image, atol=1e-10)
"""
import numpy as np
from affine_transform import transform
from mgen import rotation_from_angle
import matplotlib.pyplot as plt
# Create a simple white square in an image
original = np.zeros((601, 401))
original[100:300, 100:300] = 1
# Rotate by 22.5° (around the centre of the square (200,200)) and shift +200 in x and +100 in y
transformed = transform(original,
rotation_from_angle(np.pi / 8),
np.array([200, 100]),
origin=(200, 200))
#
# Matplotlib stuff only
#
plt.rcParams['savefig.facecolor'] = "#fcfcfc"
plt.figure(figsize=(8, 3.7))
ax = plt.subplot(121)
plt.imshow(original.T, origin="lower", cmap="gray", vmin=0.0, vmax=1.0)
plt.plot([0, 600], [200, 200], color=(0.161, 0.427, 0.859))
plt.plot([200, 200], [0, 400], color=(0.161, 0.427, 0.859))
ax.title.set_text("Original")
plt.xlabel("x-axis")
plt.ylabel("y-axis")
plt.tight_layout()
def test_smallest_image():
for dim in range(1, 6):
image = np.ones((1, ) * dim)
output = transform(image, np.eye(dim), translation=(0, ) * dim)
assert output == 1
def test_wrong_order():
image = np.ones((1, ))
with pytest.raises(ValueError):
transform(image, np.eye(1), translation=(0, ), order="fantastic")
def test_different_input_argument_types():
image = np.ones((1, ), dtype=int)
transform(image, np.eye(1, dtype=int), translation=(0, ))
def test_identity_transform():
for image_size in range(2, 6):
for dim in range(1, 6):
image = np.ones((image_size, ) * dim)
output = transform(image, np.eye(dim), translation=(0, ) * dim)
np.testing.assert_allclose(output, image)
def test_non_square_shape():
for image_size_x, image_size_y in ((x, 6 - x) for x in range(2, 6)):
image = np.ones((image_size_x, image_size_y))
output = transform(image, np.eye(2), translation=(0, 0))
np.testing.assert_allclose(output, image)
