source : nibabel-like image object
Source image
target : nibabel-like image
Target image array
similarity : str or callable
Cost-function for assessing image similarity. If a string, one
of ‘cc’, ‘cr’, ‘crl1’, ‘mi’, je’, ‘ce’, ‘nmi’, ‘smi’. ‘cr’
(correlation ratio) is the default. If a callable, it should
take a two-dimensional array representing the image joint
histogram as an input and return a float. See
registration_module.pyx
interp : str
Interpolation method. One of ‘pv’: Partial volume, ‘tri’:
Trilinear, ‘rand’: Random interpolation. See
iconic.c
subsampling : None or sequence length 3
subsampling of image in voxels, where None (default) results
in the subsampling to be automatically adjusted to roughly match
a cubic grid of 64**3 voxels
search : str or sequence
If a string, one of ‘affine’, ‘rigid’, ‘similarity’; default ‘affine’
A sequence of strings can be provided to run a graduate search, e.g.
by doing first ‘rigid’, then ‘similarity’, then ‘affine’
optimizer : str or sequence
If a string, one of ‘simplex’, ‘powell’, ‘steepest’, ‘cg’, ‘bfgs’
Alternatively, a sequence of such strings can be provided to
run several optimizers sequentially. If bot search and
optimizer are sequences, then the shorter is filled with its
last value to match the longer.
|
