def test_identity_resample():
""" Test resampling with an identity affine.
"""
shape = (3, 2, 5, 2)
data = np.random.randint(0, 10, shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine, interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
# Smoke-test with a list affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine.tolist(),
interpolation='nearest')
# Test with a 3x3 affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine[:3, :3],
interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
# Test with non native endian data
# Test with big endian data ('>f8')
for interpolation in ['nearest', 'continuous']:
rot_img = resample_img(Nifti1Image(data.astype('>f8'), affine),
target_affine=affine.tolist(),
interpolation=interpolation)
np.testing.assert_almost_equal(data, rot_img.get_data())
# Test with little endian data ('<f8')
for interpolation in ['nearest', 'continuous']:
rot_img = resample_img(Nifti1Image(data.astype('<f8'), affine),
target_affine=affine.tolist(),
interpolation=interpolation)
np.testing.assert_almost_equal(data, rot_img.get_data())
def test_resampling_continuous_with_affine():
prng = np.random.RandomState(10)
data_3d = prng.randint(1, 4, size=(1, 10, 10))
data_4d = prng.randint(1, 4, size=(1, 10, 10, 3))
for data in [data_3d, data_4d]:
for angle in (0, np.pi / 2., np.pi, 3 * np.pi / 2.):
rot = rotation(0, angle)
img = Nifti1Image(data, np.eye(4))
rot_img = resample_img(
img,
target_affine=rot,
interpolation='continuous')
rot_img_back = resample_img(
rot_img,
target_affine=np.eye(4),
interpolation='continuous')
center = slice(1, 9)
# values on the edges are wrong for some reason
mask = (0, center, center)
np.testing.assert_allclose(
img.get_data()[mask],
rot_img_back.get_data()[mask])
assert_equal(rot_img.get_data().dtype,
np.dtype(data.dtype.name.replace('int', 'float')))
def test_resampling_with_affine():
""" Test resampling with a given rotation part of the affine.
"""
prng = np.random.RandomState(10)
data_3d = prng.randint(4, size=(1, 4, 4))
data_4d = prng.randint(4, size=(1, 4, 4, 3))
for data in [data_3d, data_4d]:
for angle in (0, np.pi, np.pi / 2., np.pi / 4., np.pi / 3.):
rot = rotation(0, angle)
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest')
assert_equal(np.max(data),
np.max(rot_img.get_data()))
assert_equal(rot_img.get_data().dtype, data.dtype)
# We take the same rotation logic as above and test with nonnative endian
# data as input
for data in [data_3d, data_4d]:
img = Nifti1Image(data.astype('>f8'), np.eye(4))
for angle in (0, np.pi, np.pi / 2., np.pi / 4., np.pi / 3.):
rot = rotation(0, angle)
rot_img = resample_img(img, target_affine=rot,
interpolation='nearest')
assert_equal(np.max(data),
np.max(rot_img.get_data()))
def save_maps(model_dir, doc, resample=False,
target_affine=None, target_shape=None):
for dtype in ['c_maps', 't_maps']:
if dtype in doc:
maps_dir = make_dir(model_dir, dtype, strict=False)
for key in doc[dtype]:
fname = '%s.nii.gz' % safe_name(key.lower())
img = nb.load(doc[dtype][key])
if resample:
img = resample_img(img, target_affine, target_shape)
nb.save(img, os.path.join(maps_dir, fname))
if 'beta_maps' in doc:
maps_dir = make_dir(model_dir, 'beta_maps')
for path in doc['beta_maps']:
fname = '%s.nii.gz' % safe_name(os.path.split(
path)[1].lower().split('.')[0])
img = nb.load(path)
if resample:
img = resample_img(
img, target_affine, target_shape, copy=False)
nb.save(img, os.path.join(maps_dir, fname))
if 'mask' in doc:
img = nb.load(doc['mask'])
if resample:
img = resample_img(img, target_affine, target_shape,
interpolation='nearest', copy=False)
nb.save(img, os.path.join(model_dir, 'mask.nii.gz'))
def test_resampling_fill_value():
""" Test resampling with a non-zero fill value
"""
prng = np.random.RandomState(10)
data_3d = prng.rand(1, 4, 4)
data_4d = prng.rand(1, 4, 4, 3)
angle = np.pi/4
rot = rotation(0, angle)
# Try a few different fill values
for data in [data_3d, data_4d]:
for val in (-3.75, 0):
if val:
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest',
fill_value=val,
clip=False)
else:
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest',
clip=False)
assert_equal(rot_img.get_data().flatten()[0],
val)
rot_img2 = resample_to_img(Nifti1Image(data, np.eye(4)),
rot_img,
interpolation='nearest',
fill_value=val)
assert_equal(rot_img2.get_data().flatten()[0],
val)
def test_resampling_with_int_types_no_crash():
affine = np.eye(4)
data = np.zeros((2, 2, 2))
for dtype in [np.int, np.int8, np.int16, np.int32, np.int64,
np.uint, np.uint8, np.uint16, np.uint32, np.uint64,
np.float32, np.float64, np.float]:
img = Nifti1Image(data.astype(dtype), affine)
resample_img(img, target_affine=2. * affine)
def test_resampling_nan():
# Test that when the data has NaNs they do not propagate to the
# whole image
for core_shape in [(3, 5, 4), (3, 5, 4, 2)]:
# create deterministic data, padded with one
# voxel thickness of zeros
core_data = np.arange(np.prod(core_shape)
).reshape(core_shape).astype(np.float)
# Introduce a nan
core_data[2, 2:4, 1] = np.nan
full_data_shape = np.array(core_shape) + 2
full_data = np.zeros(full_data_shape)
full_data[[slice(1, 1 + s) for s in core_shape]] = core_data
source_img = Nifti1Image(full_data, np.eye(4))
# Transform real data using easily checkable transformations
# For now: axis permutations
axis_permutation = [0, 1, 2]
# check 3x3 transformation matrix
target_affine = np.eye(3)[axis_permutation]
with pytest.warns(Warning, match=r"(\bnan\b|invalid value)"):
resampled_img = resample_img(source_img,
target_affine=target_affine)
resampled_data = get_data(resampled_img)
if full_data.ndim == 4:
axis_permutation.append(3)
what_resampled_data_should_be = full_data.transpose(axis_permutation)
non_nan = np.isfinite(what_resampled_data_should_be)
# Check that the input data hasn't been modified:
assert not np.all(non_nan)
# Check that for finite value resampling works without problems
assert_array_almost_equal(resampled_data[non_nan],
what_resampled_data_should_be[non_nan])
# Check that what was not finite is still not finite
assert not np.any(np.isfinite(
resampled_data[np.logical_not(non_nan)]))
# Test with an actual resampling, in the case of a bigish hole
# This checks the extrapolation mechanism: if we don't do any
# extrapolation before resampling, the hole creates big
# artefacts
data = 10 * np.ones((10, 10, 10))
data[4:6, 4:6, 4:6] = np.nan
source_img = Nifti1Image(data, 2 * np.eye(4))
with pytest.warns(RuntimeWarning):
resampled_img = resample_img(source_img,
target_affine=np.eye(4))
resampled_data = get_data(resampled_img)
np.testing.assert_allclose(10, resampled_data[np.isfinite(resampled_data)])
def test_resampling_with_int_types_no_crash():
affine = np.eye(4)
data = np.zeros((2, 2, 2))
for dtype in [
np.int, np.int8, np.int16, np.int32, np.int64, np.uint, np.uint8,
np.uint16, np.uint32, np.uint64, np.float32, np.float64, np.float
]:
img = Nifti1Image(data.astype(dtype), affine)
resample_img(img, target_affine=2. * affine)
def test_raises_upon_3x3_affine_and_no_shape():
img = Nifti1Image(np.zeros([8, 9, 10]), affine=np.eye(4))
exception = ValueError
message = ("Given target shape without anchor "
"vector: Affine shape should be \\(4, 4\\) and "
"not \\(3, 3\\)")
with pytest.raises(exception, match=message):
resample_img(img,
target_affine=np.eye(3) * 2,
target_shape=(10, 10, 10))
def test_4d_affine_bounding_box_error():
small_data = np.ones([4, 4, 4])
small_data_4D_affine = np.eye(4)
small_data_4D_affine[:3, -1] = np.array([5, 4, 5])
small_img = Nifti1Image(small_data,
small_data_4D_affine)
bigger_data_4D_affine = np.eye(4)
bigger_data = np.zeros([10, 10, 10])
bigger_img = Nifti1Image(bigger_data,
bigger_data_4D_affine)
# We would like to check whether all/most of the data
# will be contained in the resampled image
# The measure will be the l2 norm, since some resampling
# schemes approximately conserve it
def l2_norm(arr):
return (arr ** 2).sum()
# resample using 4D affine and specified target shape
small_to_big_with_shape = resample_img(
small_img,
target_affine=bigger_img.affine,
target_shape=bigger_img.shape)
# resample using 3D affine and no target shape
small_to_big_without_shape_3D_affine = resample_img(
small_img,
target_affine=bigger_img.affine[:3, :3])
# resample using 4D affine and no target shape
small_to_big_without_shape = resample_img(
small_img,
target_affine=bigger_img.affine)
# The first 2 should pass
assert_almost_equal(l2_norm(small_data),
l2_norm(get_data(small_to_big_with_shape)))
assert_almost_equal(
l2_norm(small_data),
l2_norm(get_data(small_to_big_without_shape_3D_affine)))
# After correcting decision tree for 4x4 affine given + no target shape
# from "use initial shape" to "calculate minimal bounding box respecting
# the affine anchor and the data"
assert_almost_equal(l2_norm(small_data),
l2_norm(get_data(small_to_big_without_shape)))
assert_array_equal(
small_to_big_without_shape.shape,
small_data_4D_affine[:3, -1] + np.array(small_img.shape))
def test_resample_img_segmentation_fault():
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
img_in = Nifti1Image(data, aff_in)
resample_img(img_in, target_affine=aff_out, interpolation='nearest')
def test_4d_affine_bounding_box_error():
small_data = np.ones([4, 4, 4])
small_data_4D_affine = np.eye(4)
small_data_4D_affine[:3, -1] = np.array([5, 4, 5])
small_img = Nifti1Image(small_data,
small_data_4D_affine)
bigger_data_4D_affine = np.eye(4)
bigger_data = np.zeros([10, 10, 10])
bigger_img = Nifti1Image(bigger_data,
bigger_data_4D_affine)
# We would like to check whether all/most of the data
# will be contained in the resampled image
# The measure will be the l2 norm, since some resampling
# schemes approximately conserve it
def l2_norm(arr):
return (arr ** 2).sum()
# resample using 4D affine and specified target shape
small_to_big_with_shape = resample_img(
small_img,
target_affine=bigger_img.get_affine(),
target_shape=bigger_img.shape)
# resample using 3D affine and no target shape
small_to_big_without_shape_3D_affine = resample_img(
small_img,
target_affine=bigger_img.get_affine()[:3, :3])
# resample using 4D affine and no target shape
small_to_big_without_shape = resample_img(
small_img,
target_affine=bigger_img.get_affine())
# The first 2 should pass
assert_almost_equal(l2_norm(small_data),
l2_norm(small_to_big_with_shape.get_data()))
assert_almost_equal(l2_norm(small_data),
l2_norm(small_to_big_without_shape_3D_affine.get_data()))
# After correcting decision tree for 4x4 affine given + no target shape
# from "use initial shape" to "calculate minimal bounding box respecting
# the affine anchor and the data"
assert_almost_equal(l2_norm(small_data),
l2_norm(small_to_big_without_shape.get_data()))
assert_array_equal(small_to_big_without_shape.shape,
small_data_4D_affine[:3, -1] + np.array(small_img.shape))
def test_resampling_error_checks():
shape = (3, 2, 5, 2)
target_shape = (5, 3, 2)
affine = np.eye(4)
data = np.random.randint(0, 10, shape)
img = Nifti1Image(data, affine)
# Correct parameters: no exception
resample_img(img, target_shape=target_shape, target_affine=affine)
resample_img(img, target_affine=affine)
with testing.write_tmp_imgs(img) as filename:
resample_img(filename, target_shape=target_shape, target_affine=affine)
# Missing parameter
assert_raises(ValueError, resample_img, img, target_shape=target_shape)
# Invalid shape
assert_raises(ValueError, resample_img, img, target_shape=(2, 3),
target_affine=affine)
# Invalid interpolation
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
testing.assert_raises_regex(ValueError, pattern,
resample_img, img, target_shape=target_shape,
target_affine=affine,
interpolation="an_invalid_interpolation")
# Noop
target_shape = shape[:3]
img_r = resample_img(img, copy=False)
assert_equal(img_r, img)
img_r = resample_img(img, copy=True)
assert_false(np.may_share_memory(img_r.get_data(), img.get_data()))
np.testing.assert_almost_equal(img_r.get_data(), img.get_data())
np.testing.assert_almost_equal(img_r.get_affine(), img.get_affine())
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=False)
assert_equal(img_r, img)
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=True)
assert_false(np.may_share_memory(img_r.get_data(), img.get_data()))
np.testing.assert_almost_equal(img_r.get_data(), img.get_data())
np.testing.assert_almost_equal(img_r.get_affine(), img.get_affine())
def test_resampling_error_checks():
shape = (3, 2, 5, 2)
target_shape = (5, 3, 2)
affine = np.eye(4)
data = np.random.randint(0, 10, shape)
img = Nifti1Image(data, affine)
# Correct parameters: no exception
resample_img(img, target_shape=target_shape, target_affine=affine)
resample_img(img, target_affine=affine)
with testing.write_tmp_imgs(img) as filename:
resample_img(filename, target_shape=target_shape, target_affine=affine)
# Missing parameter
assert_raises(ValueError, resample_img, img, target_shape=target_shape)
# Invalid shape
assert_raises(ValueError, resample_img, img, target_shape=(2, 3),
target_affine=affine)
# Invalid interpolation
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
testing.assert_raises_regex(ValueError, pattern,
resample_img, img, target_shape=target_shape,
target_affine=affine,
interpolation="an_invalid_interpolation")
# Noop
target_shape = shape[:3]
img_r = resample_img(img, copy=False)
assert_equal(img_r, img)
img_r = resample_img(img, copy=True)
assert_false(np.may_share_memory(img_r.get_data(), img.get_data()))
np.testing.assert_almost_equal(img_r.get_data(), img.get_data())
np.testing.assert_almost_equal(img_r.affine, img.affine)
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=False)
assert_equal(img_r, img)
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=True)
assert_false(np.may_share_memory(img_r.get_data(), img.get_data()))
np.testing.assert_almost_equal(img_r.get_data(), img.get_data())
np.testing.assert_almost_equal(img_r.affine, img.affine)
def test_resample_img_segmentation_fault():
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
img_in = Nifti1Image(data, aff_in)
resample_img(img_in,
target_affine=aff_out,
interpolation='nearest')
def linear_downsampling_generator(generator,
max_downsampling_factor=2,
isotropic=False):
'''
Downsamples each sample (linearly) by a random factor and upsamples to original resolution again (nearest neighbor).
Info:
* Uses nilearn resample_img for resampling.
* If isotropic=True: Resamples all dimensions (channels, x, y, z) with same downsampling factor
* If isotropic=False: Randomly choose new downsampling factor for each dimension
* Does not resample "seg".
'''
import nibabel as nib
from nilearn.image.resampling import resample_img, resample_to_img
for data_dict in generator:
assert "data" in list(
data_dict.keys()
), "your data generator needs to return a python dictionary with at least a 'data' key value pair"
data = data_dict[
'data'] # shape of data must be: (batch_size, nr_of_channels, x, y, [z]) (z ist optional; nr_of_channels can be 1)
dim = len(
data.shape[2:]) # remove batch_size and nr_of_channels dimension
for sample_idx in range(data.shape[0]):
fact = random.uniform(1, max_downsampling_factor)
for channel_idx in range(data.shape[1]):
affine = np.identity(4)
if dim == 3:
img_data = data[sample_idx, channel_idx]
elif dim == 2:
tmp = data[sample_idx, channel_idx]
img_data = np.reshape(
tmp, (1, tmp.shape[0], tmp.shape[1])
) # add third spatial dimension to make resample_img work
else:
raise ValueError("Invalid dimension size")
image = nib.Nifti1Image(img_data, affine)
affine2 = affine
if isotropic:
affine2[0, 0] = fact
affine2[1, 1] = fact
affine2[2, 2] = fact
else:
affine2[0, 0] = random.uniform(1, max_downsampling_factor)
affine2[1, 1] = random.uniform(1, max_downsampling_factor)
affine2[2, 2] = random.uniform(1, max_downsampling_factor)
affine2[3, 3] = 1
image2 = resample_img(image,
target_affine=affine2,
interpolation='continuous')
image3 = resample_to_img(image2, image, 'nearest')
data[sample_idx, channel_idx] = np.squeeze(image3.get_data())
data_dict["data"] = data
yield data_dict
def test_resampling_result_axis_permutation():
# Transform real data using easily checkable transformations
# For now: axis permutations
# create a cuboid full of deterministic data, padded with one
# voxel thickness of zeros
core_shape = (3, 5, 4)
core_data = np.arange(np.prod(core_shape)).reshape(core_shape)
full_data_shape = np.array(core_shape) + 2
full_data = np.zeros(full_data_shape)
full_data[[slice(1, 1 + s) for s in core_shape]] = core_data
source_img = Nifti1Image(full_data, np.eye(4))
axis_permutations = [[0, 1, 2],
[1, 0, 2],
[2, 1, 0],
[0, 2, 1]]
# check 3x3 transformation matrix
for ap in axis_permutations:
target_affine = np.eye(3)[ap]
resampled_img = resample_img(source_img,
target_affine=target_affine)
resampled_data = resampled_img.get_data()
what_resampled_data_should_be = full_data.transpose(ap)
assert_array_almost_equal(resampled_data,
what_resampled_data_should_be)
# check 4x4 transformation matrix
offset = np.array([-2, 1, -3])
for ap in axis_permutations:
target_affine = np.eye(4)
target_affine[:3, :3] = np.eye(3)[ap]
target_affine[:3, 3] = offset
resampled_img = resample_img(source_img,
target_affine=target_affine)
resampled_data = resampled_img.get_data()
offset_cropping = np.vstack([-offset[ap][np.newaxis, :],
np.zeros([1, 3])]
).T.ravel().astype(int)
what_resampled_data_should_be = pad(full_data.transpose(ap),
*list(offset_cropping))
assert_array_almost_equal(resampled_data,
what_resampled_data_should_be)
def test_resampling_result_axis_permutation():
# Transform real data using easily checkable transformations
# For now: axis permutations
# create a cuboid full of deterministic data, padded with one
# voxel thickness of zeros
core_shape = (3, 5, 4)
core_data = np.arange(np.prod(core_shape)).reshape(core_shape)
full_data_shape = np.array(core_shape) + 2
full_data = np.zeros(full_data_shape)
full_data[[slice(1, 1 + s) for s in core_shape]] = core_data
source_img = Nifti1Image(full_data, np.eye(4))
axis_permutations = [[0, 1, 2],
[1, 0, 2],
[2, 1, 0],
[0, 2, 1]]
# check 3x3 transformation matrix
for ap in axis_permutations:
target_affine = np.eye(3)[ap]
resampled_img = resample_img(source_img,
target_affine=target_affine)
resampled_data = resampled_img.get_data()
what_resampled_data_should_be = full_data.transpose(ap)
assert_array_almost_equal(resampled_data,
what_resampled_data_should_be)
# check 4x4 transformation matrix
offset = np.array([-2, 1, -3])
for ap in axis_permutations:
target_affine = np.eye(4)
target_affine[:3, :3] = np.eye(3)[ap]
target_affine[:3, 3] = offset
resampled_img = resample_img(source_img,
target_affine=target_affine)
resampled_data = resampled_img.get_data()
offset_cropping = np.vstack([-offset[ap][np.newaxis, :],
np.zeros([1, 3])]
).T.ravel().astype(int)
what_resampled_data_should_be = pad(full_data.transpose(ap),
*list(offset_cropping))
assert_array_almost_equal(resampled_data,
what_resampled_data_should_be)
def resample(input_image,
target_affine,
target_shape,
interpolation='continuous'):
outim = resample_img(input_image,
target_affine=target_affine,
target_shape=target_shape,
interpolation=interpolation)
return outim
def test_resample_img_segmentation_fault():
if 'APPVEYOR_TEST' in os.environ:
raise(SkipTest('Skipped on appveyor due to not enough memory on image'))
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
img_in = Nifti1Image(data, aff_in)
resample_img(img_in,
target_affine=aff_out,
interpolation='nearest')
def test_resample_img_segmentation_fault():
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
try:
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
except MemoryError:
# This can happen on AppVeyor and for 32-bit Python on Windows
pytest.skip('Not enough RAM to run this test')
else:
img_in = Nifti1Image(data, aff_in)
resample_img(img_in, target_affine=aff_out, interpolation='nearest')
def test_identity_resample():
""" Test resampling with an identity affine.
"""
shape = (3, 2, 5, 2)
data = np.random.randint(0, 10, shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine, interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
# Smoke-test with a list affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine.tolist(),
interpolation='nearest')
# Test with a 3x3 affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine[:3, :3],
interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
def test_raises_bbox_error_if_data_outside_box():
# Make some cases which should raise exceptions
# original image
data = np.zeros([8, 9, 10])
affine = np.eye(4)
affine_offset = np.array([1, 1, 1])
affine[:3, 3] = affine_offset
img = Nifti1Image(data, affine)
# some axis flipping affines
axis_flips = np.array(list(map(np.diag,
[[-1, 1, 1, 1],
[1, -1, 1, 1],
[1, 1, -1, 1],
[-1, -1, 1, 1],
[-1, 1, -1, 1],
[1, -1, -1, 1]])))
# some in plane 90 degree rotations base on these
# (by permuting two lines)
af = axis_flips
rotations = np.array([af[0][[1, 0, 2, 3]],
af[0][[2, 1, 0, 3]],
af[1][[1, 0, 2, 3]],
af[1][[0, 2, 1, 3]],
af[2][[2, 1, 0, 3]],
af[2][[0, 2, 1, 3]]])
new_affines = np.concatenate([axis_flips,
rotations])
new_offset = np.array([0., 0., 0.])
new_affines[:, :3, 3] = new_offset[np.newaxis, :]
for new_affine in new_affines:
exception = BoundingBoxError
message = ("The field of view given "
"by the target affine does "
"not contain any of the data")
with pytest.raises(exception, match=message):
resample_img(img, target_affine=new_affine)
def _resample_img(path,
target_affine=None,
target_shape=None,
interpolation='continuous'):
img = resample_img(path,
target_affine,
target_shape,
interpolation,
copy=False)
nb.save(img, path)
def test_downsample():
""" Test resampling with a 1/2 down-sampling affine.
"""
rand_gen = np.random.RandomState(0)
shape = (6, 3, 6, 2)
data = rand_gen.random_sample(shape)
affine = np.eye(4)
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=2 * affine,
interpolation='nearest')
downsampled = data[::2, ::2, ::2, ...]
x, y, z = downsampled.shape[:3]
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
rot_img_2 = resample_img(Nifti1Image(data, affine),
target_affine=2 * affine,
interpolation='nearest',
force_resample=True)
np.testing.assert_almost_equal(rot_img_2.get_data(), rot_img.get_data())
# Test with non native endian data
# Test to check that if giving non native endian data as input should
# work as normal and expected to return the same output as above tests.
# Big endian data ('>f8')
for copy in [True, False]:
rot_img = resample_img(Nifti1Image(data.astype('>f8'), affine),
target_affine=2 * affine,
interpolation='nearest',
copy=copy)
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
# Little endian data
for copy in [True, False]:
rot_img = resample_img(Nifti1Image(data.astype('<f8'), affine),
target_affine=2 * affine,
interpolation='nearest',
copy=copy)
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
def test_identity_resample():
""" Test resampling with an identity affine.
"""
shape = (3, 2, 5, 2)
data = np.random.randint(0, 10, shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine,
interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
# Smoke-test with a list affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine.tolist(),
interpolation='nearest')
# Test with a 3x3 affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine[:3, :3],
interpolation='nearest')
np.testing.assert_almost_equal(data, rot_img.get_data())
def test_resample_img_segmentation_fault():
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
try:
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
except MemoryError:
# This can happen on AppVeyor and for 32-bit Python on Windows
raise SkipTest('Not enough RAM to run this test')
img_in = Nifti1Image(data, aff_in)
resample_img(img_in,
target_affine=aff_out,
interpolation='nearest')
def test_resampling_with_affine():
""" Test resampling with a given rotation part of the affine.
"""
prng = np.random.RandomState(10)
data = prng.randint(4, size=(1, 4, 4))
for angle in (0, np.pi, np.pi / 2, np.pi / 4, np.pi / 3):
rot = rotation(0, angle)
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest')
np.testing.assert_almost_equal(np.max(data),
np.max(rot_img.get_data()))
def test_resample_img_segmentation_fault():
if os.environ.get('APPVEYOR') == 'True':
raise SkipTest('This test too slow (7-8 minutes) on AppVeyor')
# see https://github.com/nilearn/nilearn/issues/346
shape_in = (64, 64, 64)
aff_in = np.diag([2., 2., 2., 1.])
aff_out = np.diag([3., 3., 3., 1.])
# fourth_dim = 1024 works fine but for 1025 creates a segmentation
# fault with scipy < 0.14.1
fourth_dim = 1025
try:
data = np.ones(shape_in + (fourth_dim, ), dtype=np.float64)
except MemoryError:
# This can happen on AppVeyor and for 32-bit Python on Windows
raise SkipTest('Not enough RAM to run this test')
img_in = Nifti1Image(data, aff_in)
resample_img(img_in, target_affine=aff_out, interpolation='nearest')
def test_resampling_with_affine():
""" Test resampling with a given rotation part of the affine.
"""
prng = np.random.RandomState(10)
data = prng.randint(4, size=(1, 4, 4))
for angle in (0, np.pi, np.pi / 2., np.pi / 4., np.pi / 3.):
rot = rotation(0, angle)
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest')
np.testing.assert_almost_equal(np.max(data),
np.max(rot_img.get_data()))
def linear_downsampling_generator(generator, max_downsampling_factor=2, isotropic=False):
'''
Downsamples each sample (linearly) by a random factor and upsamples to original resolution again (nearest neighbor).
Info:
* Uses nilearn resample_img for resampling.
* If isotropic=True: Resamples all dimensions (channels, x, y, z) with same downsampling factor
* If isotropic=False: Randomly choose new downsampling factor for each dimension
* Does not resample "seg".
'''
import nibabel as nib
from nilearn.image.resampling import resample_img, resample_to_img
for data_dict in generator:
assert "data" in list(
data_dict.keys()), "your data generator needs to return a python dictionary with at least a 'data' key value pair"
data = data_dict[
'data'] # shape of data must be: (batch_size, nr_of_channels, x, y, [z]) (z ist optional; nr_of_channels can be 1)
dim = len(data.shape[2:]) # remove batch_size and nr_of_channels dimension
for sample_idx in range(data.shape[0]):
fact = random.uniform(1, max_downsampling_factor)
for channel_idx in range(data.shape[1]):
affine = np.identity(4)
if dim == 3:
img_data = data[sample_idx, channel_idx]
elif dim == 2:
tmp = data[sample_idx, channel_idx]
img_data = np.reshape(tmp, (
1, tmp.shape[0], tmp.shape[1])) # add third spatial dimension to make resample_img work
else:
raise ValueError("Invalid dimension size")
image = nib.Nifti1Image(img_data, affine)
affine2 = affine
if isotropic:
affine2[0, 0] = fact
affine2[1, 1] = fact
affine2[2, 2] = fact
else:
affine2[0, 0] = random.uniform(1, max_downsampling_factor)
affine2[1, 1] = random.uniform(1, max_downsampling_factor)
affine2[2, 2] = random.uniform(1, max_downsampling_factor)
affine2[3, 3] = 1
image2 = resample_img(image, target_affine=affine2, interpolation='continuous')
image3 = resample_to_img(image2, image, 'nearest')
data[sample_idx, channel_idx] = np.squeeze(image3.get_data())
data_dict["data"] = data
yield data_dict
def test_downsample():
""" Test resampling with a 1/2 down-sampling affine.
"""
rand_gen = np.random.RandomState(0)
shape = (6, 3, 6, 2)
data = rand_gen.random_sample(shape)
affine = np.eye(4)
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=2 * affine, interpolation='nearest')
downsampled = data[::2, ::2, ::2, ...]
x, y, z = downsampled.shape[:3]
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
def augment_linear_downsampling_nilearn(data,
max_downsampling_factor=2,
isotropic=False):
'''
Downsamples each sample (linearly) by a random factor and upsamples to original resolution again (nearest neighbor).
Info:
* Uses nilearn resample_img for resampling.
* If isotropic=True: Resamples all channels (channels, x, y, z) with same downsampling factor
* If isotropic=False: Randomly choose new downsampling factor for each dimension
'''
import nibabel as nib
from nilearn.image.resampling import resample_img, resample_to_img
dim = len(data.shape[2:]) # remove batch_size and nr_of_channels dimension
for sample_idx in range(data.shape[0]):
fact = random.uniform(1, max_downsampling_factor)
for channel_idx in range(data.shape[1]):
affine = np.identity(4)
if dim == 3:
img_data = data[sample_idx, channel_idx]
elif dim == 2:
tmp = data[sample_idx, channel_idx]
img_data = np.reshape(
tmp, (1, tmp.shape[0], tmp.shape[1])
) # add third spatial dimension to make resample_img work
else:
raise ValueError("Invalid dimension size")
image = nib.Nifti1Image(img_data, affine)
affine2 = affine
if isotropic:
affine2[0, 0] = fact
affine2[1, 1] = fact
affine2[2, 2] = fact
else:
affine2[0, 0] = random.uniform(1, max_downsampling_factor)
affine2[1, 1] = random.uniform(1, max_downsampling_factor)
affine2[2, 2] = random.uniform(1, max_downsampling_factor)
affine2[3, 3] = 1
image2 = resample_img(image,
target_affine=affine2,
interpolation='continuous')
image3 = resample_to_img(image2, image, 'nearest')
data[sample_idx, channel_idx] = np.squeeze(image3.get_data())
return data
def test_downsample():
""" Test resampling with a 1/2 down-sampling affine.
"""
rand_gen = np.random.RandomState(0)
shape = (6, 3, 6, 2)
data = rand_gen.random_sample(shape)
affine = np.eye(4)
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=2 * affine,
interpolation='nearest')
downsampled = data[::2, ::2, ::2, ...]
x, y, z = downsampled.shape[:3]
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
def test_identity_resample():
""" Test resampling with an identity affine.
"""
rng = np.random.RandomState(42)
shape = (3, 2, 5, 2)
data = rng.randint(0, 10, shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine,
interpolation='nearest')
np.testing.assert_almost_equal(data, get_data(rot_img))
# Smoke-test with a list affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine.tolist(),
interpolation='nearest')
# Test with a 3x3 affine
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=affine[:3, :3],
interpolation='nearest')
np.testing.assert_almost_equal(data, get_data(rot_img))
# Test with non native endian data
# Test with big endian data ('>f8')
for interpolation in ['nearest', 'linear', 'continuous']:
rot_img = resample_img(Nifti1Image(data.astype('>f8'), affine),
target_affine=affine.tolist(),
interpolation=interpolation)
np.testing.assert_almost_equal(data, get_data(rot_img))
# Test with little endian data ('<f8')
for interpolation in ['nearest', 'linear', 'continuous']:
rot_img = resample_img(Nifti1Image(data.astype('<f8'), affine),
target_affine=affine.tolist(),
interpolation=interpolation)
np.testing.assert_almost_equal(data, get_data(rot_img))
def test_downsample():
""" Test resampling with a 1/2 down-sampling affine.
"""
rand_gen = np.random.RandomState(0)
shape = (6, 3, 6, 2)
data = rand_gen.random_sample(shape)
affine = np.eye(4)
rot_img = resample_img(Nifti1Image(data, affine),
target_affine=2 * affine, interpolation='nearest')
downsampled = data[::2, ::2, ::2, ...]
x, y, z = downsampled.shape[:3]
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
# Test with non native endian data
# Test to check that if giving non native endian data as input should
# work as normal and expected to return the same output as above tests.
# Big endian data ('>f8')
for copy in [True, False]:
rot_img = resample_img(Nifti1Image(data.astype('>f8'), affine),
target_affine=2 * affine,
interpolation='nearest',
copy=copy)
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
# Little endian data
for copy in [True, False]:
rot_img = resample_img(Nifti1Image(data.astype('<f8'), affine),
target_affine=2 * affine,
interpolation='nearest',
copy=copy)
np.testing.assert_almost_equal(downsampled,
rot_img.get_data()[:x, :y, :z, ...])
def test_resample_clip():
# Resample and image and get larger and smaller
# value than in the original. Use clip to get rid of these images
shape = (6, 3, 6)
data = np.zeros(shape=shape)
data[1:-2, 1:-1, 1:-2] = 1
source_affine = np.diag((2, 2, 2, 1))
source_img = Nifti1Image(data, source_affine)
target_affine = np.eye(4)
no_clip_data = resample_img(source_img, target_affine,
clip=False).get_data()
clip_data = resample_img(source_img,
target_affine, clip=True).get_data()
not_clip = np.where((no_clip_data > data.min()) & (no_clip_data < data.max()))
assert_true(np.any(no_clip_data > data.max()))
assert_true(np.any(no_clip_data < data.min()))
assert_true(np.all(clip_data <= data.max()))
assert_true(np.all(clip_data >= data.min()))
assert_array_equal(no_clip_data[not_clip], clip_data[not_clip])
def _get_resampled_img(dtype):
data = np.ones((10, 10, 10), dtype=dtype)
data[3:7, 3:7, 3:7] = 2
affine = np.eye(4)
theta = math.pi / 6.
c = math.cos(theta)
s = math.sin(theta)
affine = np.array([[1, 0, 0, 0], [0, c, -s, 0], [0, s, c, 0],
[0, 0, 0, 1]])
img = Nifti1Image(data, affine)
return resample_img(img, target_affine=np.eye(4))
def test_resample_clip():
# Resample and image and get larger and smaller
# value than in the original. Use clip to get rid of these images
shape = (6, 3, 6)
data = np.zeros(shape=shape)
data[1:-2, 1:-1, 1:-2] = 1
source_affine = np.diag((2, 2, 2, 1))
source_img = Nifti1Image(data, source_affine)
target_affine = np.eye(4)
no_clip_data = resample_img(source_img, target_affine,
clip=False).get_data()
clip_data = resample_img(source_img, target_affine, clip=True).get_data()
not_clip = np.where((no_clip_data > data.min())
& (no_clip_data < data.max()))
assert_true(np.any(no_clip_data > data.max()))
assert_true(np.any(no_clip_data < data.min()))
assert_true(np.all(clip_data <= data.max()))
assert_true(np.all(clip_data >= data.min()))
assert_array_equal(no_clip_data[not_clip], clip_data[not_clip])
def resample(input_image,
target_affine,
target_shape,
interpolation='continuous'):
input_image_data = input_image.get_data()
nan_mask = numpy.isnan(input_image_data)
if numpy.any(nan_mask):
input_image_data[nan_mask] = 0.0
input_image = nibabel.Nifti1Image(input_image_data,
input_image.get_affine())
outim = resample_img(input_image,
target_affine=target_affine,
target_shape=target_shape,
interpolation=interpolation)
return outim
def test_resampling_with_affine():
""" Test resampling with a given rotation part of the affine.
"""
prng = np.random.RandomState(10)
data_3d = prng.randint(4, size=(1, 4, 4))
data_4d = prng.randint(4, size=(1, 4, 4, 3))
for data in [data_3d, data_4d]:
for angle in (0, np.pi, np.pi / 2., np.pi / 4., np.pi / 3.):
rot = rotation(0, angle)
rot_img = resample_img(Nifti1Image(data, np.eye(4)),
target_affine=rot,
interpolation='nearest')
assert_equal(np.max(data), np.max(rot_img.get_data()))
assert_equal(rot_img.get_data().dtype, data.dtype)
def test_mean_img_resample():
# Test resampling in mean_img with a permutation of the axes
rng = np.random.RandomState(42)
data = rng.rand(5, 6, 7, 40)
affine = np.diag((4, 3, 2, 1))
img = nibabel.Nifti1Image(data, affine=affine)
mean_img = nibabel.Nifti1Image(data.mean(axis=-1), affine=affine)
target_affine = affine[:, [1, 0, 2, 3]] # permutation of axes
mean_img_with_resampling = image.mean_img(img, target_affine=target_affine)
resampled_mean_image = resampling.resample_img(mean_img,
target_affine=target_affine)
assert_array_equal(resampled_mean_image.get_data(),
mean_img_with_resampling.get_data())
assert_array_equal(resampled_mean_image.get_affine(),
mean_img_with_resampling.get_affine())
assert_array_equal(mean_img_with_resampling.get_affine(), target_affine)
def _get_resampled_img(dtype):
data = np.ones((10, 10, 10), dtype=dtype)
data[3:7, 3:7, 3:7] = 2
affine = np.eye(4)
theta = math.pi / 6.
c = math.cos(theta)
s = math.sin(theta)
affine = np.array([[1, 0, 0, 0],
[0, c, -s, 0],
[0, s, c, 0],
[0, 0, 0, 1]])
img = Nifti1Image(data, affine)
return resample_img(img, target_affine=np.eye(4))
def augment_linear_downsampling_nilearn(data, max_downsampling_factor=2, isotropic=False):
'''
Downsamples each sample (linearly) by a random factor and upsamples to original resolution again (nearest neighbor).
Info:
* Uses nilearn resample_img for resampling.
* If isotropic=True: Resamples all channels (channels, x, y, z) with same downsampling factor
* If isotropic=False: Randomly choose new downsampling factor for each dimension
'''
import nibabel as nib
from nilearn.image.resampling import resample_img, resample_to_img
dim = len(data.shape[2:]) # remove batch_size and nr_of_channels dimension
for sample_idx in range(data.shape[0]):
fact = random.uniform(1, max_downsampling_factor)
for channel_idx in range(data.shape[1]):
affine = np.identity(4)
if dim == 3:
img_data = data[sample_idx, channel_idx]
elif dim == 2:
tmp = data[sample_idx, channel_idx]
img_data = np.reshape(tmp, (
1, tmp.shape[0], tmp.shape[1])) # add third spatial dimension to make resample_img work
else:
raise ValueError("Invalid dimension size")
image = nib.Nifti1Image(img_data, affine)
affine2 = affine
if isotropic:
affine2[0, 0] = fact
affine2[1, 1] = fact
affine2[2, 2] = fact
else:
affine2[0, 0] = random.uniform(1, max_downsampling_factor)
affine2[1, 1] = random.uniform(1, max_downsampling_factor)
affine2[2, 2] = random.uniform(1, max_downsampling_factor)
affine2[3, 3] = 1
image2 = resample_img(image, target_affine=affine2, interpolation='continuous')
image3 = resample_to_img(image2, image, 'nearest')
data[sample_idx, channel_idx] = np.squeeze(image3.get_data())
return data
def test_mean_img_resample():
# Test resampling in mean_img with a permutation of the axes
rng = np.random.RandomState(42)
data = rng.rand(5, 6, 7, 40)
affine = np.diag((4, 3, 2, 1))
img = nibabel.Nifti1Image(data, affine=affine)
mean_img = nibabel.Nifti1Image(data.mean(axis=-1), affine=affine)
target_affine = affine[:, [1, 0, 2, 3]] # permutation of axes
mean_img_with_resampling = image.mean_img(img,
target_affine=target_affine)
resampled_mean_image = resampling.resample_img(mean_img,
target_affine=target_affine)
assert_array_equal(resampled_mean_image.get_data(),
mean_img_with_resampling.get_data())
assert_array_equal(resampled_mean_image.get_affine(),
mean_img_with_resampling.get_affine())
assert_array_equal(mean_img_with_resampling.get_affine(), target_affine)
def test_3x3_affine_bbox():
# Test that the bounding-box is properly computed when
# transforming with a negative affine component
# This is specifically to test for a change in behavior between
# scipy < 0.18 and scipy >= 0.18, which is an interaction between
# offset and a diagonal affine
image = np.ones((20, 30))
source_affine = np.eye(4)
# Give the affine an offset
source_affine[:2, 3] = np.array([96, 64])
# We need to turn this data into a nibabel image
img = Nifti1Image(image[:, :, np.newaxis], affine=source_affine)
target_affine_3x3 = np.eye(3) * 2
# One negative axes
target_affine_3x3[1] *= -1
img_3d_affine = resample_img(img, target_affine=target_affine_3x3)
# If the bounding box is computed wrong, the image will be only
# zeros
np.testing.assert_allclose(img_3d_affine.get_data().max(), image.max())
def test_3x3_affine_bbox():
# Test that the bounding-box is properly computed when
# transforming with a negative affine component
# This is specifically to test for a change in behavior between
# scipy < 0.18 and scipy >= 0.18, which is an interaction between
# offset and a diagonal affine
image = np.ones((20, 30))
source_affine = np.eye(4)
# Give the affine an offset
source_affine[:2, 3] = np.array([96, 64])
# We need to turn this data into a nibabel image
img = Nifti1Image(image[:, :, np.newaxis], affine=source_affine)
target_affine_3x3 = np.eye(3) * 2
# One negative axes
target_affine_3x3[1] *= -1
img_3d_affine = resample_img(img, target_affine=target_affine_3x3)
# If the bounding box is computed wrong, the image will be only
# zeros
np.testing.assert_allclose(img_3d_affine.get_data().max(), image.max())
def test_resampling_error_checks():
rng = np.random.RandomState(42)
shape = (3, 2, 5, 2)
target_shape = (5, 3, 2)
affine = np.eye(4)
data = rng.randint(0, 10, shape)
img = Nifti1Image(data, affine)
# Correct parameters: no exception
resample_img(img, target_shape=target_shape, target_affine=affine)
resample_img(img, target_affine=affine)
with testing.write_tmp_imgs(img) as filename:
resample_img(filename, target_shape=target_shape, target_affine=affine)
# Missing parameter
pytest.raises(ValueError, resample_img, img, target_shape=target_shape)
# Invalid shape
pytest.raises(ValueError, resample_img, img, target_shape=(2, 3),
target_affine=affine)
# Invalid interpolation
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
with pytest.raises(ValueError, match=pattern):
resample_img(img,
target_shape=target_shape,
target_affine=affine,
interpolation="an_invalid_interpolation"
)
# Noop
target_shape = shape[:3]
img_r = resample_img(img, copy=False)
assert img_r == img
img_r = resample_img(img, copy=True)
assert not np.may_share_memory(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(img_r.affine, img.affine)
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=False)
assert img_r == img
img_r = resample_img(img, target_affine=affine, target_shape=target_shape,
copy=True)
assert not np.may_share_memory(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(img_r.affine, img.affine)
def test_reorder_img():
# We need to test on a square array, as rotation does not change
# shape, whereas reordering does.
shape = (5, 5, 5, 2, 2)
rng = np.random.RandomState(42)
data = rng.rand(*shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
ref_img = Nifti1Image(data, affine)
# Test with purely positive matrices and compare to a rotation
for theta, phi in np.random.randint(4, size=(5, 2)):
rot = rotation(theta * np.pi / 2, phi * np.pi / 2)
rot[np.abs(rot) < 0.001] = 0
rot[rot > 0.9] = 1
rot[rot < -0.9] = 1
b = 0.5 * np.array(shape[:3])
new_affine = from_matrix_vector(rot, b)
rot_img = resample_img(ref_img, target_affine=new_affine)
np.testing.assert_array_equal(rot_img.affine, new_affine)
np.testing.assert_array_equal(rot_img.get_data().shape, shape)
reordered_img = reorder_img(rot_img)
np.testing.assert_array_equal(reordered_img.affine[:3, :3], np.eye(3))
np.testing.assert_almost_equal(reordered_img.get_data(), data)
# Create a non-diagonal affine, and check that we raise a sensible
# exception
affine[1, 0] = 0.1
ref_img = Nifti1Image(data, affine)
testing.assert_raises_regex(ValueError, 'Cannot reorder the axes',
reorder_img, ref_img)
# Test that no exception is raised when resample='continuous'
reorder_img(ref_img, resample='continuous')
# Test that resample args gets passed to resample_img
interpolation = 'nearest'
reordered_img = reorder_img(ref_img, resample=interpolation)
resampled_img = resample_img(ref_img,
target_affine=reordered_img.affine,
interpolation=interpolation)
np.testing.assert_array_equal(reordered_img.get_data(),
resampled_img.get_data())
# Make sure invalid resample argument is included in the error message
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
testing.assert_raises_regex(ValueError,
pattern,
reorder_img,
ref_img,
resample=interpolation)
# Test flipping an axis
data = rng.rand(*shape)
for i in (0, 1, 2):
# Make a diagonal affine with a negative axis, and check that
# can be reordered, also vary the shape
shape = (i + 1, i + 2, 3 - i)
affine = np.eye(4)
affine[i, i] *= -1
img = Nifti1Image(data, affine)
orig_img = copy.copy(img)
#x, y, z = img.get_world_coords()
#sample = img.values_in_world(x, y, z)
img2 = reorder_img(img)
# Check that img has not been changed
np.testing.assert_array_equal(img.affine, orig_img.affine)
np.testing.assert_array_equal(img.get_data(), orig_img.get_data())
# Test that the affine is indeed diagonal:
np.testing.assert_array_equal(img2.affine[:3, :3],
np.diag(np.diag(img2.affine[:3, :3])))
assert_true(np.all(np.diag(img2.affine) >= 0))
def test_resampling_error_checks():
rng = np.random.RandomState(42)
shape = (3, 2, 5, 2)
target_shape = (5, 3, 2)
affine = np.eye(4)
data = rng.randint(0, 10, shape)
img = Nifti1Image(data, affine)
# Correct parameters: no exception
resample_img(img, target_shape=target_shape, target_affine=affine)
resample_img(img, target_affine=affine)
with testing.write_tmp_imgs(img) as filename:
resample_img(filename, target_shape=target_shape, target_affine=affine)
# Missing parameter
pytest.raises(ValueError, resample_img, img, target_shape=target_shape)
# Invalid shape
pytest.raises(ValueError,
resample_img,
img,
target_shape=(2, 3),
target_affine=affine)
# Invalid interpolation
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
with pytest.raises(ValueError, match=pattern):
resample_img(img,
target_shape=target_shape,
target_affine=affine,
interpolation="an_invalid_interpolation")
# Resampling a binary image with continuous or
# linear interpolation should raise a warning.
data_binary = rng.randint(4, size=(1, 4, 4))
data_binary[data_binary > 0] = 1
assert sorted(list(np.unique(data_binary))) == [0, 1]
rot = rotation(0, np.pi / 4)
img_binary = Nifti1Image(data_binary, np.eye(4))
assert _utils.niimg._is_binary_niimg(img_binary)
with pytest.warns(Warning, match="Resampling binary images with"):
rot_img = resample_img(img_binary,
target_affine=rot,
interpolation='continuous')
with pytest.warns(Warning, match="Resampling binary images with"):
rot_img = resample_img(img_binary,
target_affine=rot,
interpolation='linear')
# Noop
target_shape = shape[:3]
img_r = resample_img(img, copy=False)
assert img_r == img
img_r = resample_img(img, copy=True)
assert not np.may_share_memory(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(img_r.affine, img.affine)
img_r = resample_img(img,
target_affine=affine,
target_shape=target_shape,
copy=False)
assert img_r == img
img_r = resample_img(img,
target_affine=affine,
target_shape=target_shape,
copy=True)
assert not np.may_share_memory(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(get_data(img_r), get_data(img))
np.testing.assert_almost_equal(img_r.affine, img.affine)
def _resample_img(path, target_affine=None, target_shape=None,
interpolation='continuous'):
img = resample_img(path, target_affine, target_shape,
interpolation, copy=False)
nb.save(img, path)
def test_reorder_img():
# We need to test on a square array, as rotation does not change
# shape, whereas reordering does.
shape = (5, 5, 5, 2, 2)
rng = np.random.RandomState(42)
data = rng.rand(*shape)
affine = np.eye(4)
affine[:3, -1] = 0.5 * np.array(shape[:3])
ref_img = Nifti1Image(data, affine)
# Test with purely positive matrices and compare to a rotation
for theta, phi in np.random.randint(4, size=(5, 2)):
rot = rotation(theta * np.pi / 2, phi * np.pi / 2)
rot[np.abs(rot) < 0.001] = 0
rot[rot > 0.9] = 1
rot[rot < -0.9] = 1
b = 0.5 * np.array(shape[:3])
new_affine = from_matrix_vector(rot, b)
rot_img = resample_img(ref_img, target_affine=new_affine)
np.testing.assert_array_equal(rot_img.get_affine(), new_affine)
np.testing.assert_array_equal(rot_img.get_data().shape, shape)
reordered_img = reorder_img(rot_img)
np.testing.assert_array_equal(reordered_img.get_affine()[:3, :3],
np.eye(3))
np.testing.assert_almost_equal(reordered_img.get_data(),
data)
# Create a non-diagonal affine, and check that we raise a sensible
# exception
affine[1, 0] = 0.1
ref_img = Nifti1Image(data, affine)
testing.assert_raises_regex(ValueError, 'Cannot reorder the axes',
reorder_img, ref_img)
# Test that no exception is raised when resample='continuous'
reorder_img(ref_img, resample='continuous')
# Test that resample args gets passed to resample_img
interpolation = 'nearest'
reordered_img = reorder_img(ref_img, resample=interpolation)
resampled_img = resample_img(ref_img,
target_affine=reordered_img.get_affine(),
interpolation=interpolation)
np.testing.assert_array_equal(reordered_img.get_data(),
resampled_img.get_data())
# Make sure invalid resample argument is included in the error message
interpolation = 'an_invalid_interpolation'
pattern = "interpolation must be either.+{0}".format(interpolation)
testing.assert_raises_regex(ValueError, pattern,
reorder_img, ref_img,
resample=interpolation)
# Test flipping an axis
data = rng.rand(*shape)
for i in (0, 1, 2):
# Make a diagonal affine with a negative axis, and check that
# can be reordered, also vary the shape
shape = (i + 1, i + 2, 3 - i)
affine = np.eye(4)
affine[i, i] *= -1
img = Nifti1Image(data, affine)
orig_img = copy.copy(img)
#x, y, z = img.get_world_coords()
#sample = img.values_in_world(x, y, z)
img2 = reorder_img(img)
# Check that img has not been changed
np.testing.assert_array_equal(img.get_affine(),
orig_img.get_affine())
np.testing.assert_array_equal(img.get_data(),
orig_img.get_data())
# Test that the affine is indeed diagonal:
np.testing.assert_array_equal(img2.get_affine()[:3, :3],
np.diag(np.diag(
img2.get_affine()[:3, :3])))
assert_true(np.all(np.diag(img2.get_affine()) >= 0))
