def run(fixed0, fixed1, moving0, moving1, output):
""" Registers the two files and outputs the transformed moving file """
# Read images using the SimpleITK IO functions
fixed = [pydeform.read_volume(fixed0), pydeform.read_volume(fixed1)]
moving = [pydeform.read_volume(moving0), pydeform.read_volume(moving1)]
fixed_mask = build_mask(fixed[0])
moving_mask = build_mask(moving[0])
regularization_map = build_regularization_map(fixed[0], 100, 0.15, 0.25)
# Perform the registration
with open('registration_log.txt', 'w') as f:
df = pydeform.register(
fixed,
moving,
settings=settings,
log=f, # Log output
log_level=pydeform.LogLevel.Verbose, # Log everything
fixed_mask=fixed_mask,
moving_mask=moving_mask,
regularization_map=regularization_map)
# Transform with linear interpolation
transformed = pydeform.transform(moving[0], df,
pydeform.Interpolator.Linear)
# Write transformed image to file
pydeform.write_volume(output, transformed)
def test_affine(self):
# Test affine initialization
affine_transform = pydeform.AffineTransform(
np.array((
(2, 0, 0),
(0, 3, 0),
(0, 0, 4)
)),
np.array((10, 10, 10))
)
# Do a registration pass without actual iterations to see if affine transform is
# applied to the resulting displacement field
settings = {
'max_iteration_count': 0
}
fixed = pydeform.Volume(np.zeros((10,10,10), dtype=np.float32))
moving = pydeform.Volume(np.zeros((10,10,10), dtype=np.float32))
df = pydeform.register(
fixed,
moving,
settings=settings,
affine_transform=affine_transform
)
df = np.array(df, copy=False)
# Ax + b -> A(1, 1, 1) + b -> (2, 3, 4) + (10, 10, 10) -> (12, 13, 14)
# u(x) = Ax + b - x
self.assertEqual(df[1,1,1,0], 11)
self.assertEqual(df[1,1,1,1], 12)
self.assertEqual(df[1,1,1,2], 13)
def run(fixed_file, moving_file, output, affine_file=None):
""" Registers the two files and outputs the transformed moving file """
# Read images using the IO functions
fixed = pydeform.read_volume(fixed_file)
moving = pydeform.read_volume(moving_file)
# To use multiple image pairs you simply replace fixed and moving with lists
# of images. Remember to add an additional image_slot in the settings.
# fixed = [pydeform.read_volume((fixed_file0), pydeform.read_volume((fixed_file1)]
# moving = [pydeform.read_volume((moving_file0), pydeform.read_volume((moving_file1)]
# Optionally, read an affine transform
kwargs = {}
if affine_file:
kwargs['affine_transform'] = pydeform.read_affine_transform(
affine_file)
# Perform the registration
with open('registration_log.txt', 'w') as f:
df = pydeform.register(
fixed,
moving,
settings=settings,
log=f, # Log output
log_level=pydeform.LogLevel.Verbose, # Log everything
use_gpu=use_gpu,
**kwargs)
# Transform with linear interpolation
transformed = pydeform.transform(moving, df, pydeform.Interpolator.Linear)
# Write transformed image to file
pydeform.write_volume(output, transformed)
def _registration_worker(q, args, kwargs):
R""" To be ran in a subprocess.
Parameters
----------
q: multiprocessing.Queue
Queue to return the result.
kwargs: dict
Keyword arguments for the registration.
"""
try:
result = pydeform.register(*args, **kwargs)
except BaseException as e:
result = e
q.put(result)
def test_register(self):
with self.assertRaises(RuntimeError):
pydeform.register(None, None)
fixed = rand(10, 10, 10)
with self.assertRaises(RuntimeError):
pydeform.register(fixed, None)
moving = rand(10, 10, 10)
with self.assertRaises(RuntimeError):
pydeform.register(None, moving)
fixed = rand(10, 10, 10)
moving = [fixed, fixed]
with self.assertRaises(ValueError):
pydeform.register(fixed, moving)
moving = rand(10, 10, 10)
fixed = [moving, moving]
with self.assertRaises(ValueError):
pydeform.register(fixed, moving)
fixed = (_gauss3(size=(40, 50, 60), mu=(20, 25, 30), sigma=8) > 0.3).astype(np.float32)
moving = (_gauss3(size=(40, 50, 60), mu=(30, 20, 25), sigma=8) > 0.3).astype(np.float32)
settings = {
'regularization_weight': 0.05,
}
d = pydeform.register(fixed, moving, settings=settings)
res = _transform(moving, np.array(d), interpolator=sitk.sitkNearestNeighbor)
self.assertGreater(_jaccard(res > 0.1, fixed > 0.1), 0.97)