This commit is contained in:
ton
2023-10-05 00:01:27 +07:00
parent 1541297f6d
commit 4a987d90c5
12169 changed files with 502 additions and 2656459 deletions

View File

@@ -1,13 +0,0 @@
import numpy as np
from skimage.transform import frt2, ifrt2
def test_frt():
SIZE = 59 # must be prime to ensure that f inverse is unique
# Generate a test image
L = np.tri(SIZE, dtype=np.int32) + np.tri(SIZE, dtype=np.int32)[::-1]
f = frt2(L)
fi = ifrt2(f)
assert np.array_equal(L, fi)

View File

@@ -1,819 +0,0 @@
import re
import textwrap
import numpy as np
import pytest
from numpy.testing import (assert_almost_equal, assert_array_almost_equal,
assert_equal)
from skimage.transform import (AffineTransform, EssentialMatrixTransform,
EuclideanTransform, FundamentalMatrixTransform,
PiecewiseAffineTransform, PolynomialTransform,
ProjectiveTransform, SimilarityTransform,
estimate_transform, matrix_transform)
from skimage.transform._geometric import (GeometricTransform,
_affine_matrix_from_vector,
_center_and_normalize_points,
_euler_rotation_matrix)
SRC = np.array([
[-12.3705, -10.5075],
[-10.7865, 15.4305],
[8.6985, 10.8675],
[11.4975, -9.5715],
[7.8435, 7.4835],
[-5.3325, 6.5025],
[6.7905, -6.3765],
[-6.1695, -0.8235],
])
DST = np.array([
[0, 0],
[0, 5800],
[4900, 5800],
[4900, 0],
[4479, 4580],
[1176, 3660],
[3754, 790],
[1024, 1931],
])
def test_estimate_transform():
for tform in ('euclidean', 'similarity', 'affine', 'projective',
'polynomial'):
estimate_transform(tform, SRC[:2, :], DST[:2, :])
with pytest.raises(ValueError):
estimate_transform('foobar', SRC[:2, :], DST[:2, :])
def test_matrix_transform():
tform = AffineTransform(scale=(0.1, 0.5), rotation=2)
assert_equal(tform(SRC), matrix_transform(SRC, tform.params))
def test_euclidean_estimation():
# exact solution
tform = estimate_transform('euclidean', SRC[:2, :], SRC[:2, :] + 10)
assert_almost_equal(tform(SRC[:2, :]), SRC[:2, :] + 10)
assert_almost_equal(tform.params[0, 0], tform.params[1, 1])
assert_almost_equal(tform.params[0, 1], - tform.params[1, 0])
# over-determined
tform2 = estimate_transform('euclidean', SRC, DST)
assert_almost_equal(tform2.inverse(tform2(SRC)), SRC)
assert_almost_equal(tform2.params[0, 0], tform2.params[1, 1])
assert_almost_equal(tform2.params[0, 1], - tform2.params[1, 0])
# via estimate method
tform3 = EuclideanTransform()
assert tform3.estimate(SRC, DST)
assert_almost_equal(tform3.params, tform2.params)
def test_3d_euclidean_estimation():
src_points = np.random.rand(1000, 3)
# Random transformation for testing
angles = np.random.random((3,)) * 2 * np.pi - np.pi
rotation_matrix = _euler_rotation_matrix(angles)
translation_vector = np.random.random((3,))
dst_points = []
for pt in src_points:
pt_r = pt.reshape(3, 1)
dst = np.matmul(rotation_matrix, pt_r) + \
translation_vector.reshape(3, 1)
dst = dst.reshape(3)
dst_points.append(dst)
dst_points = np.array(dst_points)
# estimating the transformation
tform = EuclideanTransform(dimensionality=3)
assert tform.estimate(src_points, dst_points)
estimated_rotation = tform.rotation
estimated_translation = tform.translation
assert_almost_equal(estimated_rotation, rotation_matrix)
assert_almost_equal(estimated_translation, translation_vector)
def test_euclidean_init():
# init with implicit parameters
rotation = 1
translation = (1, 1)
tform = EuclideanTransform(rotation=rotation, translation=translation)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
# init with transformation matrix
tform2 = EuclideanTransform(tform.params)
assert_almost_equal(tform2.rotation, rotation)
assert_almost_equal(tform2.translation, translation)
# test special case for scale if rotation=0
rotation = 0
translation = (1, 1)
tform = EuclideanTransform(rotation=rotation, translation=translation)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
# test special case for scale if rotation=90deg
rotation = np.pi / 2
translation = (1, 1)
tform = EuclideanTransform(rotation=rotation, translation=translation)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
def test_similarity_estimation():
# exact solution
tform = estimate_transform('similarity', SRC[:2, :], DST[:2, :])
assert_almost_equal(tform(SRC[:2, :]), DST[:2, :])
assert_almost_equal(tform.params[0, 0], tform.params[1, 1])
assert_almost_equal(tform.params[0, 1], - tform.params[1, 0])
# over-determined
tform2 = estimate_transform('similarity', SRC, DST)
assert_almost_equal(tform2.inverse(tform2(SRC)), SRC)
assert_almost_equal(tform2.params[0, 0], tform2.params[1, 1])
assert_almost_equal(tform2.params[0, 1], - tform2.params[1, 0])
# via estimate method
tform3 = SimilarityTransform()
assert tform3.estimate(SRC, DST)
assert_almost_equal(tform3.params, tform2.params)
def test_3d_similarity_estimation():
src_points = np.random.rand(1000, 3)
# Random transformation for testing
angles = np.random.random((3,)) * 2 * np.pi - np.pi
scale = np.random.randint(0, 20)
rotation_matrix = _euler_rotation_matrix(angles) * scale
translation_vector = np.random.random((3,))
dst_points = []
for pt in src_points:
pt_r = pt.reshape(3, 1)
dst = np.matmul(rotation_matrix, pt_r) + \
translation_vector.reshape(3, 1)
dst = dst.reshape(3)
dst_points.append(dst)
dst_points = np.array(dst_points)
# estimating the transformation
tform = SimilarityTransform(dimensionality=3)
assert tform.estimate(src_points, dst_points)
estimated_rotation = tform.rotation
estimated_translation = tform.translation
estimated_scale = tform.scale
assert_almost_equal(estimated_translation, translation_vector)
assert_almost_equal(estimated_scale, scale)
assert_almost_equal(estimated_rotation, rotation_matrix)
def test_similarity_init():
# init with implicit parameters
scale = 0.1
rotation = 1
translation = (1, 1)
tform = SimilarityTransform(scale=scale, rotation=rotation,
translation=translation)
assert_almost_equal(tform.scale, scale)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
# init with transformation matrix
tform2 = SimilarityTransform(tform.params)
assert_almost_equal(tform2.scale, scale)
assert_almost_equal(tform2.rotation, rotation)
assert_almost_equal(tform2.translation, translation)
# test special case for scale if rotation=0
scale = 0.1
rotation = 0
translation = (1, 1)
tform = SimilarityTransform(scale=scale, rotation=rotation,
translation=translation)
assert_almost_equal(tform.scale, scale)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
# test special case for scale if rotation=90deg
scale = 0.1
rotation = np.pi / 2
translation = (1, 1)
tform = SimilarityTransform(scale=scale, rotation=rotation,
translation=translation)
assert_almost_equal(tform.scale, scale)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
# test special case for scale where the rotation isn't exactly 90deg,
# but very close
scale = 1.0
rotation = np.pi / 2
translation = (0, 0)
params = np.array([[0, -1, 1.33226763e-15],
[1, 2.22044605e-16, -1.33226763e-15],
[0, 0, 1]])
tform = SimilarityTransform(params)
assert_almost_equal(tform.scale, scale)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.translation, translation)
def test_affine_estimation():
# exact solution
tform = estimate_transform('affine', SRC[:3, :], DST[:3, :])
assert_almost_equal(tform(SRC[:3, :]), DST[:3, :])
# over-determined
tform2 = estimate_transform('affine', SRC, DST)
assert_almost_equal(tform2.inverse(tform2(SRC)), SRC)
# via estimate method
tform3 = AffineTransform()
assert tform3.estimate(SRC, DST)
assert_almost_equal(tform3.params, tform2.params)
def test_affine_init():
# init with implicit parameters
scale = (0.1, 0.13)
rotation = 1
shear = 0.1
translation = (1, 1)
tform = AffineTransform(scale=scale, rotation=rotation, shear=shear,
translation=translation)
assert_almost_equal(tform.scale, scale)
assert_almost_equal(tform.rotation, rotation)
assert_almost_equal(tform.shear, shear)
assert_almost_equal(tform.translation, translation)
# init with transformation matrix
tform2 = AffineTransform(tform.params)
assert_almost_equal(tform2.scale, scale)
assert_almost_equal(tform2.rotation, rotation)
assert_almost_equal(tform2.shear, shear)
assert_almost_equal(tform2.translation, translation)
# scalar vs. tuple scale arguments
assert_almost_equal(AffineTransform(scale=0.5).scale,
AffineTransform(scale=(0.5, 0.5)).scale)
def test_piecewise_affine():
tform = PiecewiseAffineTransform()
assert tform.estimate(SRC, DST)
# make sure each single affine transform is exactly estimated
assert_almost_equal(tform(SRC), DST)
assert_almost_equal(tform.inverse(DST), SRC)
def test_fundamental_matrix_estimation():
src = np.array([1.839035, 1.924743, 0.543582, 0.375221,
0.473240, 0.142522, 0.964910, 0.598376,
0.102388, 0.140092, 15.994343, 9.622164,
0.285901, 0.430055, 0.091150, 0.254594]).reshape(-1, 2)
dst = np.array([1.002114, 1.129644, 1.521742, 1.846002,
1.084332, 0.275134, 0.293328, 0.588992,
0.839509, 0.087290, 1.779735, 1.116857,
0.878616, 0.602447, 0.642616, 1.028681]).reshape(-1, 2)
tform = estimate_transform('fundamental', src, dst)
# Reference values obtained using COLMAP SfM library.
tform_ref = np.array([[-0.217859, 0.419282, -0.0343075],
[-0.0717941, 0.0451643, 0.0216073],
[0.248062, -0.429478, 0.0221019]])
assert_almost_equal(tform.params, tform_ref, 6)
def test_fundamental_matrix_residuals():
essential_matrix_tform = EssentialMatrixTransform(
rotation=np.eye(3), translation=np.array([1, 0, 0]))
tform = FundamentalMatrixTransform()
tform.params = essential_matrix_tform.params
src = np.array([[0, 0], [0, 0], [0, 0]])
dst = np.array([[2, 0], [2, 1], [2, 2]])
assert_almost_equal(tform.residuals(src, dst)**2, [0, 0.5, 2])
@pytest.mark.parametrize('array_like_input', [False, True])
def test_fundamental_matrix_forward(array_like_input):
if array_like_input:
rotation = [[1, 0, 0], [0, 1, 0], [0, 0, 1]]
translation = (1, 0, 0)
else:
rotation = np.eye(3)
translation = np.array([1, 0, 0])
essential_matrix_tform = EssentialMatrixTransform(
rotation=rotation, translation=translation)
if array_like_input:
params = [list(p) for p in essential_matrix_tform.params]
else:
params = essential_matrix_tform.params
tform = FundamentalMatrixTransform(matrix=params)
src = np.array([[0, 0], [0, 1], [1, 1]])
assert_almost_equal(tform(src), [[0, -1, 0], [0, -1, 1], [0, -1, 1]])
def test_fundamental_matrix_inverse():
essential_matrix_tform = EssentialMatrixTransform(
rotation=np.eye(3), translation=np.array([1, 0, 0]))
tform = FundamentalMatrixTransform()
tform.params = essential_matrix_tform.params
src = np.array([[0, 0], [0, 1], [1, 1]])
assert_almost_equal(tform.inverse(src),
[[0, 1, 0], [0, 1, -1], [0, 1, -1]])
def test_essential_matrix_init():
tform = EssentialMatrixTransform(rotation=np.eye(3),
translation=np.array([0, 0, 1]))
assert_equal(tform.params,
np.array([0, -1, 0, 1, 0, 0, 0, 0, 0]).reshape(3, 3))
def test_essential_matrix_estimation():
src = np.array([1.839035, 1.924743, 0.543582, 0.375221,
0.473240, 0.142522, 0.964910, 0.598376,
0.102388, 0.140092, 15.994343, 9.622164,
0.285901, 0.430055, 0.091150, 0.254594]).reshape(-1, 2)
dst = np.array([1.002114, 1.129644, 1.521742, 1.846002,
1.084332, 0.275134, 0.293328, 0.588992,
0.839509, 0.087290, 1.779735, 1.116857,
0.878616, 0.602447, 0.642616, 1.028681]).reshape(-1, 2)
tform = estimate_transform('essential', src, dst)
# Reference values obtained using COLMAP SfM library.
tform_ref = np.array([[-0.0811666, 0.255449, -0.0478999],
[-0.192392, -0.0531675, 0.119547],
[0.177784, -0.22008, -0.015203]])
assert_almost_equal(tform.params, tform_ref, 6)
def test_essential_matrix_forward():
tform = EssentialMatrixTransform(rotation=np.eye(3),
translation=np.array([1, 0, 0]))
src = np.array([[0, 0], [0, 1], [1, 1]])
assert_almost_equal(tform(src), [[0, -1, 0], [0, -1, 1], [0, -1, 1]])
def test_essential_matrix_inverse():
tform = EssentialMatrixTransform(rotation=np.eye(3),
translation=np.array([1, 0, 0]))
src = np.array([[0, 0], [0, 1], [1, 1]])
assert_almost_equal(tform.inverse(src),
[[0, 1, 0], [0, 1, -1], [0, 1, -1]])
def test_essential_matrix_residuals():
tform = EssentialMatrixTransform(rotation=np.eye(3),
translation=np.array([1, 0, 0]))
src = np.array([[0, 0], [0, 0], [0, 0]])
dst = np.array([[2, 0], [2, 1], [2, 2]])
assert_almost_equal(tform.residuals(src, dst)**2, [0, 0.5, 2])
def test_projective_estimation():
# exact solution
tform = estimate_transform('projective', SRC[:4, :], DST[:4, :])
assert_almost_equal(tform(SRC[:4, :]), DST[:4, :])
# over-determined
tform2 = estimate_transform('projective', SRC, DST)
assert_almost_equal(tform2.inverse(tform2(SRC)), SRC)
# via estimate method
tform3 = ProjectiveTransform()
assert tform3.estimate(SRC, DST)
assert_almost_equal(tform3.params, tform2.params)
def test_projective_weighted_estimation():
# Exact solution with same points, and unity weights
tform = estimate_transform('projective', SRC[:4, :], DST[:4, :])
tform_w = estimate_transform('projective',
SRC[:4, :], DST[:4, :], np.ones(4))
assert_almost_equal(tform.params, tform_w.params)
# Over-determined solution with same points, and unity weights
tform = estimate_transform('projective', SRC, DST)
tform_w = estimate_transform('projective',
SRC, DST, np.ones(SRC.shape[0]))
assert_almost_equal(tform.params, tform_w.params)
# Repeating a point, but setting its weight small, should give nearly
# the same result.
point_weights = np.ones(SRC.shape[0] + 1)
point_weights[0] = 1.0e-15
tform1 = estimate_transform('projective', SRC, DST)
tform2 = estimate_transform('projective',
SRC[np.arange(-1, SRC.shape[0]), :],
DST[np.arange(-1, SRC.shape[0]), :],
point_weights)
assert_almost_equal(tform1.params, tform2.params, decimal=3)
@pytest.mark.parametrize('array_like_input', [False, True])
def test_projective_init(array_like_input):
tform = estimate_transform('projective', SRC, DST)
# init with transformation matrix
if array_like_input:
params = [list(p) for p in tform.params]
else:
params = tform.params
tform2 = ProjectiveTransform(params)
assert_almost_equal(tform2.params, tform.params)
def test_polynomial_estimation():
# over-determined
tform = estimate_transform('polynomial', SRC, DST, order=10)
assert_almost_equal(tform(SRC), DST, 6)
# via estimate method
tform2 = PolynomialTransform()
assert tform2.estimate(SRC, DST, order=10)
assert_almost_equal(tform2.params, tform.params)
def test_polynomial_weighted_estimation():
# Over-determined solution with same points, and unity weights
tform = estimate_transform('polynomial', SRC, DST, order=10)
tform_w = estimate_transform('polynomial',
SRC,
DST,
order=10,
weights=np.ones(SRC.shape[0]))
assert_almost_equal(tform.params, tform_w.params)
# Repeating a point, but setting its weight small, should give nearly
# the same result.
point_weights = np.ones(SRC.shape[0] + 1)
point_weights[0] = 1.0e-15
tform1 = estimate_transform('polynomial', SRC, DST, order=10)
tform2 = estimate_transform('polynomial',
SRC[np.arange(-1, SRC.shape[0]), :],
DST[np.arange(-1, SRC.shape[0]), :],
order=10,
weights=point_weights)
assert_almost_equal(tform1.params, tform2.params, decimal=4)
@pytest.mark.parametrize('array_like_input', [False, True])
def test_polynomial_init(array_like_input):
tform = estimate_transform('polynomial', SRC, DST, order=10)
# init with transformation parameters
if array_like_input:
params = [list(p) for p in tform.params]
else:
params = tform.params
tform2 = PolynomialTransform(params)
assert_almost_equal(tform2.params, tform.params)
def test_polynomial_default_order():
tform = estimate_transform('polynomial', SRC, DST)
tform2 = estimate_transform('polynomial', SRC, DST, order=2)
assert_almost_equal(tform2.params, tform.params)
def test_polynomial_inverse():
with pytest.raises(Exception):
PolynomialTransform().inverse(0)
def test_union():
tform1 = SimilarityTransform(scale=0.1, rotation=0.3)
tform2 = SimilarityTransform(scale=0.1, rotation=0.9)
tform3 = SimilarityTransform(scale=0.1 ** 2, rotation=0.3 + 0.9)
tform = tform1 + tform2
assert_almost_equal(tform.params, tform3.params)
tform1 = AffineTransform(scale=(0.1, 0.1), rotation=0.3)
tform2 = SimilarityTransform(scale=0.1, rotation=0.9)
tform3 = SimilarityTransform(scale=0.1 ** 2, rotation=0.3 + 0.9)
tform = tform1 + tform2
assert_almost_equal(tform.params, tform3.params)
assert tform.__class__ == ProjectiveTransform
tform = AffineTransform(scale=(0.1, 0.1), rotation=0.3)
assert_almost_equal((tform + tform.inverse).params, np.eye(3))
tform1 = SimilarityTransform(scale=0.1, rotation=0.3)
tform2 = SimilarityTransform(scale=0.1, rotation=0.9)
tform3 = SimilarityTransform(scale=0.1 * 1/0.1, rotation=0.3 - 0.9)
tform = tform1 + tform2.inverse
assert_almost_equal(tform.params, tform3.params)
def test_union_differing_types():
tform1 = SimilarityTransform()
tform2 = PolynomialTransform()
with pytest.raises(TypeError):
tform1.__add__(tform2)
def test_geometric_tform():
tform = GeometricTransform()
with pytest.raises(NotImplementedError):
tform(0)
with pytest.raises(NotImplementedError):
tform.inverse(0)
with pytest.raises(NotImplementedError):
tform.__add__(0)
# See gh-3926 for discussion details
for i in range(20):
# Generate random Homography
H = np.random.rand(3, 3) * 100
H[2, H[2] == 0] += np.finfo(float).eps
H /= H[2, 2]
# Craft some src coords
src = np.array([
[(H[2, 1] + 1) / -H[2, 0], 1],
[1, (H[2, 0] + 1) / -H[2, 1]],
[1, 1],
])
# Prior to gh-3926, under the above circumstances,
# destination coordinates could be returned with nan/inf values.
tform = ProjectiveTransform(H) # Construct the transform
dst = tform(src) # Obtain the dst coords
# Ensure dst coords are finite numeric values
assert(np.isfinite(dst).all())
def test_invalid_input():
with pytest.raises(ValueError):
ProjectiveTransform(np.zeros((2, 3)))
with pytest.raises(ValueError):
AffineTransform(np.zeros((2, 3)))
with pytest.raises(ValueError):
SimilarityTransform(np.zeros((2, 3)))
with pytest.raises(ValueError):
EuclideanTransform(np.zeros((2, 3)))
with pytest.raises(ValueError):
AffineTransform(matrix=np.zeros((2, 3)), scale=1)
with pytest.raises(ValueError):
SimilarityTransform(matrix=np.zeros((2, 3)), scale=1)
with pytest.raises(ValueError):
EuclideanTransform(
matrix=np.zeros((2, 3)), translation=(0, 0))
with pytest.raises(ValueError):
PolynomialTransform(np.zeros((3, 3)))
with pytest.raises(ValueError):
FundamentalMatrixTransform(matrix=np.zeros((3, 2)))
with pytest.raises(ValueError):
EssentialMatrixTransform(matrix=np.zeros((3, 2)))
with pytest.raises(ValueError):
EssentialMatrixTransform(rotation=np.zeros((3, 2)))
with pytest.raises(ValueError):
EssentialMatrixTransform(
rotation=np.zeros((3, 3)))
with pytest.raises(ValueError):
EssentialMatrixTransform(
rotation=np.eye(3))
with pytest.raises(ValueError):
EssentialMatrixTransform(rotation=np.eye(3),
translation=np.zeros((2,)))
with pytest.raises(ValueError):
EssentialMatrixTransform(rotation=np.eye(3),
translation=np.zeros((2,)))
with pytest.raises(ValueError):
EssentialMatrixTransform(
rotation=np.eye(3), translation=np.zeros((3,)))
def test_degenerate():
src = dst = np.zeros((10, 2))
tform = SimilarityTransform()
assert not tform.estimate(src, dst)
assert np.all(np.isnan(tform.params))
tform = EuclideanTransform()
assert not tform.estimate(src, dst)
assert np.all(np.isnan(tform.params))
tform = AffineTransform()
assert not tform.estimate(src, dst)
assert np.all(np.isnan(tform.params))
tform = ProjectiveTransform()
assert not tform.estimate(src, dst)
assert np.all(np.isnan(tform.params))
# See gh-3926 for discussion details
tform = ProjectiveTransform()
for i in range(20):
# Some random coordinates
src = np.random.rand(4, 2) * 100
dst = np.random.rand(4, 2) * 100
# Degenerate the case by arranging points on a single line
src[:, 1] = np.random.rand()
# Prior to gh-3926, under the above circumstances,
# a transform could be returned with nan values.
assert(not tform.estimate(src, dst) or np.isfinite(tform.params).all())
src = np.array([[0, 2, 0], [0, 2, 0], [0, 4, 0]])
dst = np.array([[0, 1, 0], [0, 1, 0], [0, 3, 0]])
tform = AffineTransform()
assert not tform.estimate(src, dst)
# Prior to gh-6207, the above would set the parameters as the identity.
assert np.all(np.isnan(tform.params))
# The tessellation on the following points produces one degenerate affine
# warp within PiecewiseAffineTransform.
src = np.asarray([
[0, 192, 256], [0, 256, 256], [5, 0, 192], [5, 64, 0], [5, 64, 64],
[5, 64, 256], [5, 192, 192], [5, 256, 256], [0, 192, 256],
])
dst = np.asarray([
[0, 142, 206], [0, 206, 206], [5, -50, 142], [5, 14, 0], [5, 14, 64],
[5, 14, 206], [5, 142, 142], [5, 206, 206], [0, 142, 206],
])
tform = PiecewiseAffineTransform()
assert not tform.estimate(src, dst)
assert np.all(np.isnan(tform.affines[4].params)) # degenerate affine
for idx, affine in enumerate(tform.affines):
if idx != 4:
assert not np.all(np.isnan(affine.params))
for affine in tform.inverse_affines:
assert not np.all(np.isnan(affine.params))
def test_normalize_degenerate_points():
"""Return nan matrix *of appropriate size* when point is repeated."""
pts = np.array([[73.42834308, 94.2977623]] * 3)
mat, pts_tf = _center_and_normalize_points(pts)
assert np.all(np.isnan(mat))
assert np.all(np.isnan(pts_tf))
assert mat.shape == (3, 3)
assert pts_tf.shape == pts.shape
def test_projective_repr():
tform = ProjectiveTransform()
want = re.escape(textwrap.dedent(
'''
<ProjectiveTransform(matrix=
[[1., 0., 0.],
[0., 1., 0.],
[0., 0., 1.]]) at
''').strip()) + ' 0x[a-f0-9]+' + re.escape('>')
# Hack the escaped regex to allow whitespace before each number for
# compatibility with different numpy versions.
want = want.replace('0\\.', ' *0\\.')
want = want.replace('1\\.', ' *1\\.')
assert re.match(want, repr(tform))
def test_projective_str():
tform = ProjectiveTransform()
want = re.escape(textwrap.dedent(
'''
<ProjectiveTransform(matrix=
[[1., 0., 0.],
[0., 1., 0.],
[0., 0., 1.]])>
''').strip())
# Hack the escaped regex to allow whitespace before each number for
# compatibility with different numpy versions.
want = want.replace('0\\.', ' *0\\.')
want = want.replace('1\\.', ' *1\\.')
print(want)
assert re.match(want, str(tform))
def _assert_least_squares(tf, src, dst):
baseline = np.sum((tf(src) - dst) ** 2)
for i in range(tf.params.size):
for update in [0.001, -0.001]:
params = np.copy(tf.params)
params.flat[i] += update
new_tf = tf.__class__(matrix=params)
new_ssq = np.sum((new_tf(src) - dst) ** 2)
assert new_ssq > baseline
@pytest.mark.parametrize('array_like_input', [False, True])
def test_estimate_affine_3d(array_like_input):
ndim = 3
src = np.random.random((25, ndim)) * 2 ** np.arange(7, 7 + ndim)
matrix = np.array([
[4.8, 0.1, 0.2, 25],
[0.0, 1.0, 0.1, 30],
[0.0, 0.0, 1.0, -2],
[0.0, 0.0, 0.0, 1.]
])
if array_like_input:
# list of lists for matrix and src coords
src = [list(c) for c in src]
matrix = [list(c) for c in matrix]
tf = AffineTransform(matrix=matrix)
dst = tf(src)
dst_noisy = dst + np.random.random((25, ndim))
if array_like_input:
# list of lists for destination coords
dst = [list(c) for c in dst]
tf2 = AffineTransform(dimensionality=ndim)
assert tf2.estimate(src, dst_noisy)
# we check rot/scale/etc more tightly than translation because translation
# estimation is on the 1 pixel scale
matrix = np.asarray(matrix)
assert_almost_equal(tf2.params[:, :-1], matrix[:, :-1], decimal=2)
assert_almost_equal(tf2.params[:, -1], matrix[:, -1], decimal=0)
_assert_least_squares(tf2, src, dst_noisy)
def test_fundamental_3d_not_implemented():
with pytest.raises(NotImplementedError):
_ = FundamentalMatrixTransform(dimensionality=3)
with pytest.raises(NotImplementedError):
_ = FundamentalMatrixTransform(np.eye(4))
def test_array_protocol():
mat = np.eye(4)
tf = ProjectiveTransform(mat)
assert_equal(np.array(tf), mat)
assert_equal(np.array(tf, dtype=int), mat.astype(int))
def test_affine_transform_from_linearized_parameters():
mat = np.concatenate(
(np.random.random((3, 4)), np.eye(4)[-1:]), axis=0
)
v = mat[:-1].ravel()
mat_from_v = _affine_matrix_from_vector(v)
tf = AffineTransform(matrix=mat_from_v)
assert_equal(np.array(tf), mat)
# incorrect number of parameters
with pytest.raises(ValueError):
_ = _affine_matrix_from_vector(v[:-1])
with pytest.raises(ValueError):
_ = AffineTransform(matrix=v[:-1])
def test_affine_params_nD_error():
with pytest.raises(ValueError):
_ = AffineTransform(scale=5, dimensionality=3)
def test_euler_rotation():
v = [0, 10, 0]
angles = np.radians([90, 45, 45])
expected = [-5, -5, 7.1]
R = _euler_rotation_matrix(angles)
assert_almost_equal(R @ v, expected, decimal=1)
def test_euclidean_param_defaults():
# 2D rotation is 0 when only translation is given
tf = EuclideanTransform(translation=(5, 5))
assert np.array(tf)[0, 1] == 0
# off diagonals are 0 when only translation is given
tf = EuclideanTransform(translation=(4, 5, 9), dimensionality=3)
assert_equal(np.array(tf)[[0, 0, 1, 1, 2, 2], [1, 2, 0, 2, 0, 1]], 0)
with pytest.raises(ValueError):
# specifying parameters for D>3 is not supported
_ = EuclideanTransform(translation=(5, 6, 7, 8), dimensionality=4)
with pytest.raises(ValueError):
# incorrect number of angles for given dimensionality
_ = EuclideanTransform(rotation=(4, 8), dimensionality=3)
# translation is 0 when rotation is given
tf = EuclideanTransform(rotation=np.pi * np.arange(3), dimensionality=3)
assert_equal(np.array(tf)[:-1, 3], 0)
def test_similarity_transform_params():
with pytest.raises(ValueError):
_ = SimilarityTransform(translation=(4, 5, 6, 7), dimensionality=4)
tf = SimilarityTransform(scale=4, dimensionality=3)
assert_equal(tf([[1, 1, 1]]), [[4, 4, 4]])
def test_euler_angle_consistency():
angles = np.random.random((3,)) * 2 * np.pi - np.pi
euclid = EuclideanTransform(rotation=angles, dimensionality=3)
similar = SimilarityTransform(rotation=angles, dimensionality=3)
assert_array_almost_equal(euclid, similar)
def test_2D_only_implementations():
with pytest.raises(NotImplementedError):
_ = PolynomialTransform(dimensionality=3)
tf = AffineTransform(dimensionality=3)
with pytest.raises(NotImplementedError):
_ = tf.rotation
with pytest.raises(NotImplementedError):
_ = tf.shear

View File

@@ -1,576 +0,0 @@
import numpy as np
import pytest
from numpy.testing import assert_almost_equal, assert_equal
from skimage import data, transform
from skimage._shared.testing import test_parallel
from skimage.draw import circle_perimeter, ellipse_perimeter, line
@test_parallel()
def test_hough_line():
# Generate a test image
img = np.zeros((100, 150), dtype=int)
rr, cc = line(60, 130, 80, 10)
img[rr, cc] = 1
out, angles, d = transform.hough_line(img)
y, x = np.where(out == out.max())
dist = d[y[0]]
theta = angles[x[0]]
assert_almost_equal(dist, 80.0, 1)
assert_almost_equal(theta, 1.41, 1)
def test_hough_line_angles():
img = np.zeros((10, 10))
img[0, 0] = 1
out, angles, d = transform.hough_line(img, np.linspace(0, 360, 10))
assert_equal(len(angles), 10)
def test_hough_line_bad_input():
img = np.zeros(100)
img[10] = 1
# Expected error, img must be 2D
with pytest.raises(ValueError):
transform.hough_line(img, np.linspace(0, 360, 10))
def test_probabilistic_hough():
# Generate a test image
img = np.zeros((100, 100), dtype=int)
for i in range(25, 75):
img[100 - i, i] = 100
img[i, i] = 100
# decrease default theta sampling because similar orientations may confuse
# as mentioned in article of Galambos et al
theta = np.linspace(0, np.pi, 45)
lines = transform.probabilistic_hough_line(
img, threshold=10, line_length=10, line_gap=1, theta=theta)
# sort the lines according to the x-axis
sorted_lines = []
for ln in lines:
ln = list(ln)
ln.sort(key=lambda x: x[0])
sorted_lines.append(ln)
assert([(25, 75), (74, 26)] in sorted_lines)
assert([(25, 25), (74, 74)] in sorted_lines)
# Execute with default theta
transform.probabilistic_hough_line(img, line_length=10, line_gap=3)
def test_probabilistic_hough_seed():
# Load image that is likely to give a randomly varying number of lines
image = data.checkerboard()
# Use constant seed to ensure a deterministic output
lines = transform.probabilistic_hough_line(image, threshold=50,
line_length=50, line_gap=1,
seed=41537233)
assert len(lines) == 56
def test_probabilistic_hough_bad_input():
img = np.zeros(100)
img[10] = 1
# Expected error, img must be 2D
with pytest.raises(ValueError):
transform.probabilistic_hough_line(img)
def test_hough_line_peaks():
img = np.zeros((100, 150), dtype=int)
rr, cc = line(60, 130, 80, 10)
img[rr, cc] = 1
out, angles, d = transform.hough_line(img)
out, theta, dist = transform.hough_line_peaks(out, angles, d)
assert_equal(len(dist), 1)
assert_almost_equal(dist[0], 81.0, 1)
assert_almost_equal(theta[0], 1.41, 1)
def test_hough_line_peaks_ordered():
# Regression test per PR #1421
testim = np.zeros((256, 64), dtype=bool)
testim[50:100, 20] = True
testim[20:225, 25] = True
testim[15:35, 50] = True
testim[1:-1, 58] = True
hough_space, angles, dists = transform.hough_line(testim)
hspace, _, _ = transform.hough_line_peaks(hough_space, angles, dists)
assert hspace[0] > hspace[1]
def test_hough_line_peaks_single_line():
# Regression test for gh-6187, gh-4129
# create an empty test image
img = np.zeros((100, 100), dtype=bool)
# draw a horizontal line into our test image
img[30, :] = 1
hough_space, angles, dist = transform.hough_line(img)
best_h_space, best_angles, best_dist = transform.hough_line_peaks(
hough_space, angles, dist
)
assert len(best_angles) == 1
assert len(best_dist) == 1
expected_angle = -np.pi / 2
expected_dist = -30
assert abs(best_angles[0] - expected_angle) < 0.01
assert abs(best_dist[0] - expected_dist) < 0.01
def test_hough_line_peaks_dist():
img = np.zeros((100, 100), dtype=bool)
img[:, 30] = True
img[:, 40] = True
hspace, angles, dists = transform.hough_line(img)
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_distance=5)[0]) == 2
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_distance=15)[0]) == 1
def test_hough_line_peaks_angle():
check_hough_line_peaks_angle()
def check_hough_line_peaks_angle():
img = np.zeros((100, 100), dtype=bool)
img[:, 0] = True
img[0, :] = True
hspace, angles, dists = transform.hough_line(img)
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=45)[0]) == 2
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=90)[0]) == 1
theta = np.linspace(0, np.pi, 100)
hspace, angles, dists = transform.hough_line(img, theta)
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=45)[0]) == 2
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=90)[0]) == 1
theta = np.linspace(np.pi / 3, 4. / 3 * np.pi, 100)
hspace, angles, dists = transform.hough_line(img, theta)
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=45)[0]) == 2
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_angle=90)[0]) == 1
def test_hough_line_peaks_num():
img = np.zeros((100, 100), dtype=bool)
img[:, 30] = True
img[:, 40] = True
hspace, angles, dists = transform.hough_line(img)
assert len(transform.hough_line_peaks(hspace, angles, dists,
min_distance=0, min_angle=0,
num_peaks=1)[0]) == 1
def test_hough_line_peaks_zero_input():
# Test to make sure empty input doesn't cause a failure
img = np.zeros((100, 100), dtype='uint8')
theta = np.linspace(0, np.pi, 100)
hspace, angles, dists = transform.hough_line(img, theta)
h, a, d = transform.hough_line_peaks(hspace, angles, dists)
assert_equal(a, np.array([]))
def test_hough_line_peaks_single_angle():
# Regression test for gh-4814
# This code snippet used to raise an IndexError
img = np.random.random((100, 100))
tested_angles = np.array([np.pi / 2])
h, theta, d = transform.hough_line(img, theta=tested_angles)
accum, angles, dists = transform.hough_line_peaks(h, theta, d, threshold=2)
@test_parallel()
def test_hough_circle():
# Prepare picture
img = np.zeros((120, 100), dtype=int)
radius = 20
x_0, y_0 = (99, 50)
y, x = circle_perimeter(y_0, x_0, radius)
img[x, y] = 1
out1 = transform.hough_circle(img, radius)
out2 = transform.hough_circle(img, [radius])
assert_equal(out1, out2)
out = transform.hough_circle(img, np.array([radius], dtype=np.intp))
assert_equal(out, out1)
x, y = np.where(out[0] == out[0].max())
assert_equal(x[0], x_0)
assert_equal(y[0], y_0)
def test_hough_circle_extended():
# Prepare picture
# The circle center is outside the image
img = np.zeros((100, 100), dtype=int)
radius = 20
x_0, y_0 = (-5, 50)
y, x = circle_perimeter(y_0, x_0, radius)
img[x[np.where(x > 0)], y[np.where(x > 0)]] = 1
out = transform.hough_circle(img, np.array([radius], dtype=np.intp),
full_output=True)
x, y = np.where(out[0] == out[0].max())
# Offset for x_0, y_0
assert_equal(x[0], x_0 + radius)
assert_equal(y[0], y_0 + radius)
def test_hough_circle_peaks():
x_0, y_0, rad_0 = (99, 50, 20)
img = np.zeros((120, 100), dtype=int)
y, x = circle_perimeter(y_0, x_0, rad_0)
img[x, y] = 1
x_1, y_1, rad_1 = (49, 60, 30)
y, x = circle_perimeter(y_1, x_1, rad_1)
img[x, y] = 1
radii = [rad_0, rad_1]
hspaces = transform.hough_circle(img, radii)
out = transform.hough_circle_peaks(hspaces, radii, min_xdistance=1,
min_ydistance=1, threshold=None,
num_peaks=np.inf,
total_num_peaks=np.inf)
s = np.argsort(out[3]) # sort by radii
assert_equal(out[1][s], np.array([y_0, y_1]))
assert_equal(out[2][s], np.array([x_0, x_1]))
assert_equal(out[3][s], np.array([rad_0, rad_1]))
def test_hough_circle_peaks_total_peak():
img = np.zeros((120, 100), dtype=int)
x_0, y_0, rad_0 = (99, 50, 20)
y, x = circle_perimeter(y_0, x_0, rad_0)
img[x, y] = 1
x_1, y_1, rad_1 = (49, 60, 30)
y, x = circle_perimeter(y_1, x_1, rad_1)
img[x, y] = 1
radii = [rad_0, rad_1]
hspaces = transform.hough_circle(img, radii)
out = transform.hough_circle_peaks(hspaces, radii, min_xdistance=1,
min_ydistance=1, threshold=None,
num_peaks=np.inf, total_num_peaks=1)
assert_equal(out[1][0], np.array([y_1, ]))
assert_equal(out[2][0], np.array([x_1, ]))
assert_equal(out[3][0], np.array([rad_1, ]))
def test_hough_circle_peaks_min_distance():
x_0, y_0, rad_0 = (50, 50, 20)
img = np.zeros((120, 100), dtype=int)
y, x = circle_perimeter(y_0, x_0, rad_0)
img[x, y] = 1
x_1, y_1, rad_1 = (60, 60, 30)
y, x = circle_perimeter(y_1, x_1, rad_1)
# Add noise and create an imperfect circle to lower the peak in Hough space
y[::2] += 1
x[::2] += 1
img[x, y] = 1
x_2, y_2, rad_2 = (70, 70, 20)
y, x = circle_perimeter(y_2, x_2, rad_2)
# Add noise and create an imperfect circle to lower the peak in Hough space
y[::2] += 1
x[::2] += 1
img[x, y] = 1
radii = [rad_0, rad_1, rad_2]
hspaces = transform.hough_circle(img, radii)
out = transform.hough_circle_peaks(hspaces, radii, min_xdistance=15,
min_ydistance=15, threshold=None,
num_peaks=np.inf,
total_num_peaks=np.inf,
normalize=True)
# The second circle is too close to the first one
# and has a weaker peak in Hough space due to imperfectness.
# Therefore it got removed.
assert_equal(out[1], np.array([y_0, y_2]))
assert_equal(out[2], np.array([x_0, x_2]))
assert_equal(out[3], np.array([rad_0, rad_2]))
def test_hough_circle_peaks_total_peak_and_min_distance():
img = np.zeros((120, 120), dtype=int)
cx = cy = [40, 50, 60, 70, 80]
radii = range(20, 30, 2)
for i in range(len(cx)):
y, x = circle_perimeter(cy[i], cx[i], radii[i])
img[x, y] = 1
hspaces = transform.hough_circle(img, radii)
out = transform.hough_circle_peaks(hspaces, radii, min_xdistance=15,
min_ydistance=15, threshold=None,
num_peaks=np.inf,
total_num_peaks=2,
normalize=True)
# 2nd (4th) circle is removed as it is close to 1st (3rd) oneself.
# 5th is removed as total_num_peaks = 2
assert_equal(out[1], np.array(cy[:4:2]))
assert_equal(out[2], np.array(cx[:4:2]))
assert_equal(out[3], np.array(radii[:4:2]))
def test_hough_circle_peaks_normalize():
x_0, y_0, rad_0 = (50, 50, 20)
img = np.zeros((120, 100), dtype=int)
y, x = circle_perimeter(y_0, x_0, rad_0)
img[x, y] = 1
x_1, y_1, rad_1 = (60, 60, 30)
y, x = circle_perimeter(y_1, x_1, rad_1)
img[x, y] = 1
radii = [rad_0, rad_1]
hspaces = transform.hough_circle(img, radii)
out = transform.hough_circle_peaks(hspaces, radii, min_xdistance=15,
min_ydistance=15, threshold=None,
num_peaks=np.inf,
total_num_peaks=np.inf,
normalize=False)
# Two perfect circles are close but the second one is bigger.
# Therefore, it is picked due to its high peak.
assert_equal(out[1], np.array([y_1]))
assert_equal(out[2], np.array([x_1]))
assert_equal(out[3], np.array([rad_1]))
def test_hough_ellipse_zero_angle():
img = np.zeros((25, 25), dtype=int)
rx = 6
ry = 8
x0 = 12
y0 = 15
angle = 0
rr, cc = ellipse_perimeter(y0, x0, ry, rx)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=9)
best = result[-1]
assert_equal(best[1], y0)
assert_equal(best[2], x0)
assert_almost_equal(best[3], ry, decimal=1)
assert_almost_equal(best[4], rx, decimal=1)
assert_equal(best[5], angle)
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_posangle1():
# ry > rx, angle in [0:pi/2]
img = np.zeros((30, 24), dtype=int)
rx = 6
ry = 12
x0 = 10
y0 = 15
angle = np.pi / 1.35
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
assert_almost_equal(best[1] / 100., y0 / 100., decimal=1)
assert_almost_equal(best[2] / 100., x0 / 100., decimal=1)
assert_almost_equal(best[3] / 10., ry / 10., decimal=1)
assert_almost_equal(best[4] / 100., rx / 100., decimal=1)
assert_almost_equal(best[5], angle, decimal=1)
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_posangle2():
# ry < rx, angle in [0:pi/2]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = np.pi / 1.35
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
assert_almost_equal(best[1] / 100., y0 / 100., decimal=1)
assert_almost_equal(best[2] / 100., x0 / 100., decimal=1)
assert_almost_equal(best[3] / 10., ry / 10., decimal=1)
assert_almost_equal(best[4] / 100., rx / 100., decimal=1)
assert_almost_equal(best[5], angle, decimal=1)
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_posangle3():
# ry < rx, angle in [pi/2:pi]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = np.pi / 1.35 + np.pi / 2.
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_posangle4():
# ry < rx, angle in [pi:3pi/4]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = np.pi / 1.35 + np.pi
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_negangle1():
# ry > rx, angle in [0:-pi/2]
img = np.zeros((30, 24), dtype=int)
rx = 6
ry = 12
x0 = 10
y0 = 15
angle = - np.pi / 1.35
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_negangle2():
# ry < rx, angle in [0:-pi/2]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = - np.pi / 1.35
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_negangle3():
# ry < rx, angle in [-pi/2:-pi]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = - np.pi / 1.35 - np.pi / 2.
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_non_zero_negangle4():
# ry < rx, angle in [-pi:-3pi/4]
img = np.zeros((30, 24), dtype=int)
rx = 12
ry = 6
x0 = 10
y0 = 15
angle = - np.pi / 1.35 - np.pi
rr, cc = ellipse_perimeter(y0, x0, ry, rx, orientation=angle)
img[rr, cc] = 1
result = transform.hough_ellipse(img, threshold=15, accuracy=3)
result.sort(order='accumulator')
best = result[-1]
# Check if I re-draw the ellipse, points are the same!
# ie check API compatibility between hough_ellipse and ellipse_perimeter
rr2, cc2 = ellipse_perimeter(y0, x0, int(best[3]), int(best[4]),
orientation=best[5])
assert_equal(rr, rr2)
assert_equal(cc, cc2)
def test_hough_ellipse_all_black_img():
assert(transform.hough_ellipse(np.zeros((100, 100))).shape == (0, 6))

View File

@@ -1,64 +0,0 @@
import numpy as np
import pytest
from numpy.testing import assert_allclose, assert_equal
from skimage.transform import integral_image, integrate
np.random.seed(0)
x = (np.random.rand(50, 50) * 255).astype(np.uint8)
s = integral_image(x)
@pytest.mark.parametrize(
'dtype', [np.float16, np.float32, np.float64, np.uint8, np.int32]
)
@pytest.mark.parametrize('dtype_as_kwarg', [False, True])
def test_integral_image_validity(dtype, dtype_as_kwarg):
rstate = np.random.default_rng(1234)
dtype_kwarg = dtype if dtype_as_kwarg else None
y = (rstate.random((20, 20)) * 255).astype(dtype)
out = integral_image(y, dtype=dtype_kwarg)
if y.dtype.kind == 'f':
if dtype_as_kwarg:
assert out.dtype == dtype
rtol = 1e-3 if dtype == np.float16 else 1e-7
assert_allclose(out[-1, -1], y.sum(dtype=np.float64), rtol=rtol)
else:
assert out.dtype == np.float64
assert_allclose(out[-1, -1], y.sum(dtype=np.float64))
else:
assert out.dtype.kind == y.dtype.kind
if not (dtype_as_kwarg and dtype == np.uint8):
# omit check for dtype=uint8 case as it will overflow
assert_equal(out[-1, -1], y.sum())
def test_integrate_basic():
assert_equal(x[12:24, 10:20].sum(), integrate(s, (12, 10), (23, 19)))
assert_equal(x[:20, :20].sum(), integrate(s, (0, 0), (19, 19)))
assert_equal(x[:20, 10:20].sum(), integrate(s, (0, 10), (19, 19)))
assert_equal(x[10:20, :20].sum(), integrate(s, (10, 0), (19, 19)))
def test_integrate_single():
assert_equal(x[0, 0], integrate(s, (0, 0), (0, 0)))
assert_equal(x[10, 10], integrate(s, (10, 10), (10, 10)))
def test_vectorized_integrate():
r0 = np.array([12, 0, 0, 10, 0, 10, 30])
c0 = np.array([10, 0, 10, 0, 0, 10, 31])
r1 = np.array([23, 19, 19, 19, 0, 10, 49])
c1 = np.array([19, 19, 19, 19, 0, 10, 49])
expected = np.array([x[12:24, 10:20].sum(),
x[:20, :20].sum(),
x[:20, 10:20].sum(),
x[10:20, :20].sum(),
x[0, 0],
x[10, 10],
x[30:, 31:].sum()])
start_pts = [(r0[i], c0[i]) for i in range(len(r0))]
end_pts = [(r1[i], c1[i]) for i in range(len(r0))]
assert_equal(expected, integrate(s, start_pts, end_pts))

View File

@@ -1,212 +0,0 @@
import math
import pytest
import numpy as np
from numpy.testing import assert_almost_equal, assert_array_equal, assert_equal
from skimage import data
from skimage._shared.utils import _supported_float_type
from skimage.transform import pyramids
image = data.astronaut()
image_gray = image[..., 0]
@pytest.mark.parametrize('channel_axis', [0, 1, -1])
def test_pyramid_reduce_rgb(channel_axis):
image = data.astronaut()
rows, cols, dim = image.shape
image = np.moveaxis(image, source=-1, destination=channel_axis)
out_ = pyramids.pyramid_reduce(image, downscale=2,
channel_axis=channel_axis)
out = np.moveaxis(out_, channel_axis, -1)
assert_array_equal(out.shape, (rows / 2, cols / 2, dim))
def test_pyramid_reduce_gray():
rows, cols = image_gray.shape
out1 = pyramids.pyramid_reduce(image_gray, downscale=2,
channel_axis=None)
assert_array_equal(out1.shape, (rows / 2, cols / 2))
assert_almost_equal(out1.ptp(), 1.0, decimal=2)
out2 = pyramids.pyramid_reduce(image_gray, downscale=2,
channel_axis=None, preserve_range=True)
assert_almost_equal(out2.ptp() / image_gray.ptp(), 1.0, decimal=2)
def test_pyramid_reduce_gray_defaults():
rows, cols = image_gray.shape
out1 = pyramids.pyramid_reduce(image_gray)
assert_array_equal(out1.shape, (rows / 2, cols / 2))
assert_almost_equal(out1.ptp(), 1.0, decimal=2)
out2 = pyramids.pyramid_reduce(image_gray, preserve_range=True)
assert_almost_equal(out2.ptp() / image_gray.ptp(), 1.0, decimal=2)
def test_pyramid_reduce_nd():
for ndim in [1, 2, 3, 4]:
img = np.random.randn(*((8, ) * ndim))
out = pyramids.pyramid_reduce(img, downscale=2,
channel_axis=None)
expected_shape = np.asarray(img.shape) / 2
assert_array_equal(out.shape, expected_shape)
@pytest.mark.parametrize('channel_axis', [0, 1, 2, -1, -2, -3])
def test_pyramid_expand_rgb(channel_axis):
image = data.astronaut()
rows, cols, dim = image.shape
image = np.moveaxis(image, source=-1, destination=channel_axis)
out = pyramids.pyramid_expand(image, upscale=2,
channel_axis=channel_axis)
expected_shape = [rows * 2, cols * 2]
expected_shape.insert(channel_axis % image.ndim, dim)
assert_array_equal(out.shape, expected_shape)
def test_pyramid_expand_gray():
rows, cols = image_gray.shape
out = pyramids.pyramid_expand(image_gray, upscale=2)
assert_array_equal(out.shape, (rows * 2, cols * 2))
def test_pyramid_expand_nd():
for ndim in [1, 2, 3, 4]:
img = np.random.randn(*((4, ) * ndim))
out = pyramids.pyramid_expand(img, upscale=2,
channel_axis=None)
expected_shape = np.asarray(img.shape) * 2
assert_array_equal(out.shape, expected_shape)
@pytest.mark.parametrize('channel_axis', [0, 1, 2, -1, -2, -3])
def test_build_gaussian_pyramid_rgb(channel_axis):
image = data.astronaut()
rows, cols, dim = image.shape
image = np.moveaxis(image, source=-1, destination=channel_axis)
pyramid = pyramids.pyramid_gaussian(image, downscale=2,
channel_axis=channel_axis)
for layer, out in enumerate(pyramid):
layer_shape = [rows / 2 ** layer, cols / 2 ** layer]
layer_shape.insert(channel_axis % image.ndim, dim)
assert out.shape == tuple(layer_shape)
def test_build_gaussian_pyramid_gray():
rows, cols = image_gray.shape
pyramid = pyramids.pyramid_gaussian(image_gray, downscale=2,
channel_axis=None)
for layer, out in enumerate(pyramid):
layer_shape = (rows / 2 ** layer, cols / 2 ** layer)
assert_array_equal(out.shape, layer_shape)
def test_build_gaussian_pyramid_gray_defaults():
rows, cols = image_gray.shape
pyramid = pyramids.pyramid_gaussian(image_gray)
for layer, out in enumerate(pyramid):
layer_shape = (rows / 2 ** layer, cols / 2 ** layer)
assert_array_equal(out.shape, layer_shape)
def test_build_gaussian_pyramid_nd():
for ndim in [1, 2, 3, 4]:
img = np.random.randn(*((8, ) * ndim))
original_shape = np.asarray(img.shape)
pyramid = pyramids.pyramid_gaussian(img, downscale=2,
channel_axis=None)
for layer, out in enumerate(pyramid):
layer_shape = original_shape / 2 ** layer
assert_array_equal(out.shape, layer_shape)
@pytest.mark.parametrize('channel_axis', [0, 1, 2, -1, -2, -3])
def test_build_laplacian_pyramid_rgb(channel_axis):
image = data.astronaut()
rows, cols, dim = image.shape
image = np.moveaxis(image, source=-1, destination=channel_axis)
pyramid = pyramids.pyramid_laplacian(image, downscale=2,
channel_axis=channel_axis)
for layer, out in enumerate(pyramid):
layer_shape = [rows / 2 ** layer, cols / 2 ** layer]
layer_shape.insert(channel_axis % image.ndim, dim)
assert out.shape == tuple(layer_shape)
def test_build_laplacian_pyramid_defaults():
rows, cols = image_gray.shape
pyramid = pyramids.pyramid_laplacian(image_gray)
for layer, out in enumerate(pyramid):
layer_shape = (rows / 2 ** layer, cols / 2 ** layer)
assert_array_equal(out.shape, layer_shape)
def test_build_laplacian_pyramid_nd():
for ndim in [1, 2, 3, 4]:
img = np.random.randn(*(16, )*ndim)
original_shape = np.asarray(img.shape)
pyramid = pyramids.pyramid_laplacian(img, downscale=2,
channel_axis=None)
for layer, out in enumerate(pyramid):
layer_shape = original_shape / 2 ** layer
assert_array_equal(out.shape, layer_shape)
@pytest.mark.parametrize('channel_axis', [0, 1, 2, -1, -2, -3])
def test_laplacian_pyramid_max_layers(channel_axis):
for downscale in [2, 3, 5, 7]:
if channel_axis is None:
shape = (32, 8)
shape_without_channels = shape
else:
shape_without_channels = (32, 8)
ndim = len(shape_without_channels) + 1
n_channels = 5
shape = list(shape_without_channels)
shape.insert(channel_axis % ndim, n_channels)
shape = tuple(shape)
img = np.ones(shape)
pyramid = pyramids.pyramid_laplacian(img, downscale=downscale,
channel_axis=channel_axis)
max_layer = math.ceil(math.log(max(shape_without_channels), downscale))
for layer, out in enumerate(pyramid):
if channel_axis is None:
out_shape_without_channels = out.shape
else:
assert out.shape[channel_axis] == n_channels
out_shape_without_channels = list(out.shape)
out_shape_without_channels.pop(channel_axis)
out_shape_without_channels = tuple(out_shape_without_channels)
if layer < max_layer:
# should not reach all axes as size 1 prior to final level
assert max(out_shape_without_channels) > 1
# total number of images is max_layer + 1
assert_equal(max_layer, layer)
# final layer should be size 1 on all axes
assert out_shape_without_channels == (1, 1)
def test_check_factor():
with pytest.raises(ValueError):
pyramids._check_factor(0.99)
with pytest.raises(ValueError):
pyramids._check_factor(- 2)
@pytest.mark.parametrize(
'dtype', ['float16', 'float32', 'float64', 'uint8', 'int64']
)
@pytest.mark.parametrize(
'pyramid_func', [pyramids.pyramid_gaussian, pyramids.pyramid_laplacian]
)
def test_pyramid_dtype_support(pyramid_func, dtype):
img = np.random.randn(32, 8).astype(dtype)
pyramid = pyramid_func(img)
float_dtype = _supported_float_type(dtype)
assert np.all([im.dtype == float_dtype for im in pyramid])

View File

@@ -1,493 +0,0 @@
import itertools
import numpy as np
import pytest
from skimage._shared._dependency_checks import has_mpl
from skimage._shared._warnings import expected_warnings
from skimage._shared.testing import test_parallel
from skimage._shared.utils import _supported_float_type, convert_to_float
from skimage.data import shepp_logan_phantom
from skimage.transform import radon, iradon, iradon_sart, rescale
PHANTOM = shepp_logan_phantom()[::2, ::2]
PHANTOM = rescale(PHANTOM, 0.5, order=1,
mode='constant', anti_aliasing=False, channel_axis=None)
def _debug_plot(original, result, sinogram=None):
from matplotlib import pyplot as plt
imkwargs = dict(cmap='gray', interpolation='nearest')
if sinogram is None:
plt.figure(figsize=(15, 6))
sp = 130
else:
plt.figure(figsize=(11, 11))
sp = 221
plt.subplot(sp + 0)
plt.imshow(sinogram, aspect='auto', **imkwargs)
plt.subplot(sp + 1)
plt.imshow(original, **imkwargs)
plt.subplot(sp + 2)
plt.imshow(result, vmin=original.min(), vmax=original.max(), **imkwargs)
plt.subplot(sp + 3)
plt.imshow(result - original, **imkwargs)
plt.colorbar()
plt.show()
def _rescale_intensity(x):
x = x.astype(float)
x -= x.min()
x /= x.max()
return x
def test_iradon_bias_circular_phantom():
"""
test that a uniform circular phantom has a small reconstruction bias
"""
pixels = 128
xy = np.arange(-pixels / 2, pixels / 2) + 0.5
x, y = np.meshgrid(xy, xy)
image = x**2 + y**2 <= (pixels/4)**2
theta = np.linspace(0., 180., max(image.shape), endpoint=False)
sinogram = radon(image, theta=theta)
reconstruction_fbp = iradon(sinogram, theta=theta)
error = reconstruction_fbp - image
tol = 5e-5
roi_err = np.abs(np.mean(error))
assert roi_err < tol
def check_radon_center(shape, circle, dtype, preserve_range):
# Create a test image with only a single non-zero pixel at the origin
image = np.zeros(shape, dtype=dtype)
image[(shape[0] // 2, shape[1] // 2)] = 1.
# Calculate the sinogram
theta = np.linspace(0., 180., max(shape), endpoint=False)
sinogram = radon(image, theta=theta, circle=circle,
preserve_range=preserve_range)
assert sinogram.dtype == _supported_float_type(sinogram.dtype)
# The sinogram should be a straight, horizontal line
sinogram_max = np.argmax(sinogram, axis=0)
print(sinogram_max)
assert np.std(sinogram_max) < 1e-6
@pytest.mark.parametrize("shape", [(16, 16), (17, 17)])
@pytest.mark.parametrize("circle", [False, True])
@pytest.mark.parametrize(
"dtype", [np.float64, np.float32, np.float16, np.uint8, bool]
)
@pytest.mark.parametrize("preserve_range", [False, True])
def test_radon_center(shape, circle, dtype, preserve_range):
check_radon_center(shape, circle, dtype, preserve_range)
@pytest.mark.parametrize("shape", [(32, 16), (33, 17)])
@pytest.mark.parametrize("circle", [False])
@pytest.mark.parametrize("dtype", [np.float64, np.float32, np.uint8, bool])
@pytest.mark.parametrize("preserve_range", [False, True])
def test_radon_center_rectangular(shape, circle, dtype, preserve_range):
check_radon_center(shape, circle, dtype, preserve_range)
def check_iradon_center(size, theta, circle):
debug = False
# Create a test sinogram corresponding to a single projection
# with a single non-zero pixel at the rotation center
if circle:
sinogram = np.zeros((size, 1), dtype=float)
sinogram[size // 2, 0] = 1.
else:
diagonal = int(np.ceil(np.sqrt(2) * size))
sinogram = np.zeros((diagonal, 1), dtype=float)
sinogram[sinogram.shape[0] // 2, 0] = 1.
maxpoint = np.unravel_index(np.argmax(sinogram), sinogram.shape)
print('shape of generated sinogram', sinogram.shape)
print('maximum in generated sinogram', maxpoint)
# Compare reconstructions for theta=angle and theta=angle + 180;
# these should be exactly equal
reconstruction = iradon(sinogram, theta=[theta], circle=circle)
reconstruction_opposite = iradon(sinogram, theta=[theta + 180],
circle=circle)
print('rms deviance:',
np.sqrt(np.mean((reconstruction_opposite - reconstruction)**2)))
if debug and has_mpl:
import matplotlib.pyplot as plt
imkwargs = dict(cmap='gray', interpolation='nearest')
plt.figure()
plt.subplot(221)
plt.imshow(sinogram, **imkwargs)
plt.subplot(222)
plt.imshow(reconstruction_opposite - reconstruction, **imkwargs)
plt.subplot(223)
plt.imshow(reconstruction, **imkwargs)
plt.subplot(224)
plt.imshow(reconstruction_opposite, **imkwargs)
plt.show()
assert np.allclose(reconstruction, reconstruction_opposite)
sizes_for_test_iradon_center = [16, 17]
thetas_for_test_iradon_center = [0, 90]
circles_for_test_iradon_center = [False, True]
@pytest.mark.parametrize(
"size, theta, circle",
itertools.product(sizes_for_test_iradon_center,
thetas_for_test_iradon_center,
circles_for_test_iradon_center)
)
def test_iradon_center(size, theta, circle):
check_iradon_center(size, theta, circle)
def check_radon_iradon(interpolation_type, filter_type):
debug = False
image = PHANTOM
reconstructed = iradon(radon(image, circle=False), filter_name=filter_type,
interpolation=interpolation_type, circle=False)
delta = np.mean(np.abs(image - reconstructed))
print('\n\tmean error:', delta)
if debug and has_mpl:
_debug_plot(image, reconstructed)
if filter_type in ('ramp', 'shepp-logan'):
if interpolation_type == 'nearest':
allowed_delta = 0.03
else:
allowed_delta = 0.025
else:
allowed_delta = 0.05
assert delta < allowed_delta
filter_types = ["ramp", "shepp-logan", "cosine", "hamming", "hann"]
interpolation_types = ['linear', 'nearest']
radon_iradon_inputs = list(itertools.product(interpolation_types,
filter_types))
# cubic interpolation is slow; only run one test for it
radon_iradon_inputs.append(('cubic', 'shepp-logan'))
@pytest.mark.parametrize(
"interpolation_type, filter_type", radon_iradon_inputs
)
def test_radon_iradon(interpolation_type, filter_type):
check_radon_iradon(interpolation_type, filter_type)
def test_iradon_angles():
"""
Test with different number of projections
"""
size = 100
# Synthetic data
image = np.tri(size) + np.tri(size)[::-1]
# Large number of projections: a good quality is expected
nb_angles = 200
theta = np.linspace(0, 180, nb_angles, endpoint=False)
radon_image_200 = radon(image, theta=theta, circle=False)
reconstructed = iradon(radon_image_200, circle=False)
delta_200 = np.mean(abs(_rescale_intensity(image) -
_rescale_intensity(reconstructed)))
assert delta_200 < 0.03
# Lower number of projections
nb_angles = 80
radon_image_80 = radon(image, theta=theta, circle=False)
# Test whether the sum of all projections is approximately the same
s = radon_image_80.sum(axis=0)
assert np.allclose(s, s[0], rtol=0.01)
reconstructed = iradon(radon_image_80, circle=False)
delta_80 = np.mean(abs(image / np.max(image) -
reconstructed / np.max(reconstructed)))
# Loss of quality when the number of projections is reduced
assert delta_80 > delta_200
def check_radon_iradon_minimal(shape, slices):
debug = False
theta = np.arange(180)
image = np.zeros(shape, dtype=float)
image[slices] = 1.
sinogram = radon(image, theta, circle=False)
reconstructed = iradon(sinogram, theta, circle=False)
print('\n\tMaximum deviation:', np.max(np.abs(image - reconstructed)))
if debug and has_mpl:
_debug_plot(image, reconstructed, sinogram)
if image.sum() == 1:
assert (np.unravel_index(np.argmax(reconstructed), image.shape)
== np.unravel_index(np.argmax(image), image.shape))
shapes = [(3, 3), (4, 4), (5, 5)]
def generate_test_data_for_radon_iradon_minimal(shapes):
def shape2coordinates(shape):
c0, c1 = shape[0] // 2, shape[1] // 2
coordinates = itertools.product((c0 - 1, c0, c0 + 1),
(c1 - 1, c1, c1 + 1))
return coordinates
def shape2shapeandcoordinates(shape):
return itertools.product([shape], shape2coordinates(shape))
return itertools.chain.from_iterable([shape2shapeandcoordinates(shape)
for shape in shapes])
@pytest.mark.parametrize(
"shape, coordinate",
generate_test_data_for_radon_iradon_minimal(shapes)
)
def test_radon_iradon_minimal(shape, coordinate):
check_radon_iradon_minimal(shape, coordinate)
def test_reconstruct_with_wrong_angles():
a = np.zeros((3, 3))
p = radon(a, theta=[0, 1, 2], circle=False)
iradon(p, theta=[0, 1, 2], circle=False)
with pytest.raises(ValueError):
iradon(p, theta=[0, 1, 2, 3])
def _random_circle(shape):
# Synthetic random data, zero outside reconstruction circle
np.random.seed(98312871)
image = np.random.rand(*shape)
c0, c1 = np.ogrid[0:shape[0], 0:shape[1]]
r = np.sqrt((c0 - shape[0] // 2)**2 + (c1 - shape[1] // 2)**2)
radius = min(shape) // 2
image[r > radius] = 0.
return image
def test_radon_circle():
a = np.ones((10, 10))
with expected_warnings(['reconstruction circle']):
radon(a, circle=True)
# Synthetic data, circular symmetry
shape = (61, 79)
c0, c1 = np.ogrid[0:shape[0], 0:shape[1]]
r = np.sqrt((c0 - shape[0] // 2)**2 + (c1 - shape[1] // 2)**2)
radius = min(shape) // 2
image = np.clip(radius - r, 0, np.inf)
image = _rescale_intensity(image)
angles = np.linspace(0, 180, min(shape), endpoint=False)
sinogram = radon(image, theta=angles, circle=True)
assert np.all(sinogram.std(axis=1) < 1e-2)
# Synthetic data, random
image = _random_circle(shape)
sinogram = radon(image, theta=angles, circle=True)
mass = sinogram.sum(axis=0)
average_mass = mass.mean()
relative_error = np.abs(mass - average_mass) / average_mass
print(relative_error.max(), relative_error.mean())
assert np.all(relative_error < 3.2e-3)
def check_sinogram_circle_to_square(size):
from skimage.transform.radon_transform import _sinogram_circle_to_square
image = _random_circle((size, size))
theta = np.linspace(0., 180., size, False)
sinogram_circle = radon(image, theta, circle=True)
def argmax_shape(a):
return np.unravel_index(np.argmax(a), a.shape)
print('\n\targmax of circle:', argmax_shape(sinogram_circle))
sinogram_square = radon(image, theta, circle=False)
print('\targmax of square:', argmax_shape(sinogram_square))
sinogram_circle_to_square = _sinogram_circle_to_square(sinogram_circle)
print('\targmax of circle to square:',
argmax_shape(sinogram_circle_to_square))
error = abs(sinogram_square - sinogram_circle_to_square)
print(np.mean(error), np.max(error))
assert (argmax_shape(sinogram_square) ==
argmax_shape(sinogram_circle_to_square))
@pytest.mark.parametrize("size", (50, 51))
def test_sinogram_circle_to_square(size):
check_sinogram_circle_to_square(size)
def check_radon_iradon_circle(interpolation, shape, output_size):
# Forward and inverse radon on synthetic data
image = _random_circle(shape)
radius = min(shape) // 2
sinogram_rectangle = radon(image, circle=False)
reconstruction_rectangle = iradon(sinogram_rectangle,
output_size=output_size,
interpolation=interpolation,
circle=False)
sinogram_circle = radon(image, circle=True)
reconstruction_circle = iradon(sinogram_circle,
output_size=output_size,
interpolation=interpolation,
circle=True)
# Crop rectangular reconstruction to match circle=True reconstruction
width = reconstruction_circle.shape[0]
excess = int(np.ceil((reconstruction_rectangle.shape[0] - width) / 2))
s = np.s_[excess:width + excess, excess:width + excess]
reconstruction_rectangle = reconstruction_rectangle[s]
# Find the reconstruction circle, set reconstruction to zero outside
c0, c1 = np.ogrid[0:width, 0:width]
r = np.sqrt((c0 - width // 2)**2 + (c1 - width // 2)**2)
reconstruction_rectangle[r > radius] = 0.
print(reconstruction_circle.shape)
print(reconstruction_rectangle.shape)
np.allclose(reconstruction_rectangle, reconstruction_circle)
# if adding more shapes to test data, you might want to look at commit d0f2bac3f
shapes_radon_iradon_circle = ((61, 79), )
interpolations = ('nearest', 'linear')
output_sizes = (None,
min(shapes_radon_iradon_circle[0]),
max(shapes_radon_iradon_circle[0]),
97)
@pytest.mark.parametrize(
"shape, interpolation, output_size",
itertools.product(shapes_radon_iradon_circle, interpolations, output_sizes)
)
def test_radon_iradon_circle(shape, interpolation, output_size):
check_radon_iradon_circle(interpolation, shape, output_size)
def test_order_angles_golden_ratio():
from skimage.transform.radon_transform import order_angles_golden_ratio
np.random.seed(1231)
lengths = [1, 4, 10, 180]
for l in lengths:
theta_ordered = np.linspace(0, 180, l, endpoint=False)
theta_random = np.random.uniform(0, 180, l)
for theta in (theta_random, theta_ordered):
indices = [x for x in order_angles_golden_ratio(theta)]
# no duplicate indices allowed
assert len(indices) == len(set(indices))
@test_parallel()
def test_iradon_sart():
debug = False
image = rescale(PHANTOM, 0.8, mode='reflect',
channel_axis=None, anti_aliasing=False)
theta_ordered = np.linspace(0., 180., image.shape[0], endpoint=False)
theta_missing_wedge = np.linspace(0., 150., image.shape[0], endpoint=True)
for theta, error_factor in ((theta_ordered, 1.),
(theta_missing_wedge, 2.)):
sinogram = radon(image, theta, circle=True)
reconstructed = iradon_sart(sinogram, theta)
if debug and has_mpl:
from matplotlib import pyplot as plt
plt.figure()
plt.subplot(221)
plt.imshow(image, interpolation='nearest')
plt.subplot(222)
plt.imshow(sinogram, interpolation='nearest')
plt.subplot(223)
plt.imshow(reconstructed, interpolation='nearest')
plt.subplot(224)
plt.imshow(reconstructed - image, interpolation='nearest')
plt.show()
delta = np.mean(np.abs(reconstructed - image))
print('delta (1 iteration) =', delta)
assert delta < 0.02 * error_factor
reconstructed = iradon_sart(sinogram, theta, reconstructed)
delta = np.mean(np.abs(reconstructed - image))
print('delta (2 iterations) =', delta)
assert delta < 0.014 * error_factor
reconstructed = iradon_sart(sinogram, theta, clip=(0, 1))
delta = np.mean(np.abs(reconstructed - image))
print('delta (1 iteration, clip) =', delta)
assert delta < 0.018 * error_factor
np.random.seed(1239867)
shifts = np.random.uniform(-3, 3, sinogram.shape[1])
x = np.arange(sinogram.shape[0])
sinogram_shifted = np.vstack([np.interp(x + shifts[i], x,
sinogram[:, i])
for i in range(sinogram.shape[1])]).T
reconstructed = iradon_sart(sinogram_shifted, theta,
projection_shifts=shifts)
if debug and has_mpl:
from matplotlib import pyplot as plt
plt.figure()
plt.subplot(221)
plt.imshow(image, interpolation='nearest')
plt.subplot(222)
plt.imshow(sinogram_shifted, interpolation='nearest')
plt.subplot(223)
plt.imshow(reconstructed, interpolation='nearest')
plt.subplot(224)
plt.imshow(reconstructed - image, interpolation='nearest')
plt.show()
delta = np.mean(np.abs(reconstructed - image))
print('delta (1 iteration, shifted sinogram) =', delta)
assert delta < 0.022 * error_factor
@pytest.mark.parametrize("preserve_range", [True, False])
def test_iradon_dtype(preserve_range):
sinogram = np.zeros((16, 1), dtype=int)
sinogram[8, 0] = 1.
sinogram64 = sinogram.astype('float64')
sinogram32 = sinogram.astype('float32')
assert iradon(sinogram, theta=[0],
preserve_range=preserve_range).dtype == 'float64'
assert iradon(sinogram64, theta=[0],
preserve_range=preserve_range).dtype == sinogram64.dtype
assert iradon(sinogram32, theta=[0],
preserve_range=preserve_range).dtype == sinogram32.dtype
def test_radon_dtype():
img = convert_to_float(PHANTOM, False)
img32 = img.astype(np.float32)
assert radon(img).dtype == img.dtype
assert radon(img32).dtype == img32.dtype
@pytest.mark.parametrize("dtype", [np.float32, np.float64])
def test_iradon_sart_dtype(dtype):
sinogram = np.zeros((16, 1), dtype=int)
sinogram[8, 0] = 1.
sinogram64 = sinogram.astype('float64')
sinogram32 = sinogram.astype('float32')
with expected_warnings(['Input data is cast to float']):
assert iradon_sart(sinogram, theta=[0]).dtype == 'float64'
assert iradon_sart(sinogram64, theta=[0]).dtype == sinogram64.dtype
assert iradon_sart(sinogram32, theta=[0]).dtype == sinogram32.dtype
assert iradon_sart(sinogram, theta=[0], dtype=dtype).dtype == dtype
assert iradon_sart(sinogram32, theta=[0], dtype=dtype).dtype == dtype
assert iradon_sart(sinogram64, theta=[0], dtype=dtype).dtype == dtype
def test_iradon_sart_wrong_dtype():
sinogram = np.zeros((16, 1))
with pytest.raises(ValueError):
iradon_sart(sinogram, dtype=int)

File diff suppressed because it is too large Load Diff