interfaces.image¶
Reorient¶
Conform an image to a given orientation
Flips and reorder the image data array so that the axes match the
directions indicated in orientation
.
The default RAS
orientation corresponds to the first axis being ordered
from left to right, the second axis from posterior to anterior, and the
third axis from inferior to superior.
For oblique images, the original orientation is considered to be the closest plumb orientation.
No resampling is performed, and thus the output image is not de-obliqued or registered to any other image or template.
The effective transform is calculated from the original affine matrix to the reoriented affine matrix.
Examples¶
If an image is not reoriented, the original file is not modified
>>> import numpy as np
>>> from nipype.interfaces.image import Reorient
>>> reorient = Reorient(orientation='LPS')
>>> reorient.inputs.in_file = 'segmentation0.nii.gz'
>>> res = reorient.run()
>>> res.outputs.out_file
'segmentation0.nii.gz'
>>> print_affine(np.loadtxt(res.outputs.transform))
1. 0. 0. 0.
0. 1. 0. 0.
0. 0. 1. 0.
0. 0. 0. 1.
>>> reorient.inputs.orientation = 'RAS'
>>> res = reorient.run()
>>> res.outputs.out_file # doctest: +ELLIPSIS
'.../segmentation0_ras.nii.gz'
>>> print_affine(np.loadtxt(res.outputs.transform))
-1. 0. 0. 60.
0. -1. 0. 72.
0. 0. 1. 0.
0. 0. 0. 1.
Inputs:
[Mandatory]
in_file: (an existing file name)
Input image
[Optional]
orientation: ('RAS' or 'RAI' or 'RPS' or 'RPI' or 'LAS' or 'LAI' or
'LPS' or 'LPI' or 'RSA' or 'RSP' or 'RIA' or 'RIP' or 'LSA' or
'LSP' or 'LIA' or 'LIP' or 'ARS' or 'ARI' or 'ALS' or 'ALI' or
'PRS' or 'PRI' or 'PLS' or 'PLI' or 'ASR' or 'ASL' or 'AIR' or
'AIL' or 'PSR' or 'PSL' or 'PIR' or 'PIL' or 'SRA' or 'SRP' or
'SLA' or 'SLP' or 'IRA' or 'IRP' or 'ILA' or 'ILP' or 'SAR' or
'SAL' or 'SPR' or 'SPL' or 'IAR' or 'IAL' or 'IPR' or 'IPL', nipype
default value: RAS)
Target axis orientation
Outputs:
out_file: (an existing file name)
Reoriented image
transform: (an existing file name)
Affine transform from input orientation to output
Rescale¶
Rescale an image
Rescales the non-zero portion of in_file
to match the bounds of the
non-zero portion of ref_file
.
Reference values in the input and reference images are defined by the
percentile
parameter, and the reference values in each image are
identified and the remaining values are scaled accordingly.
In the case of percentile == 0
, the reference values are the maxima
and minima of each image.
If the invert
parameter is set, the input file is inverted prior to
rescaling.
Examples¶
To use a high-resolution T1w image as a registration target for a T2* image, it may be useful to invert the T1w image and rescale to the T2* range. Using the 1st and 99th percentiles may reduce the impact of outlier voxels.
>>> from nipype.interfaces.image import Rescale
>>> invert_t1w = Rescale(invert=True)
>>> invert_t1w.inputs.in_file = 'structural.nii'
>>> invert_t1w.inputs.ref_file = 'functional.nii'
>>> invert_t1w.inputs.percentile = 1.
>>> res = invert_t1w.run() # doctest: +SKIP
Inputs:
[Mandatory]
in_file: (an existing file name)
Skull-stripped image to rescale
ref_file: (an existing file name)
Skull-stripped reference image
[Optional]
invert: (a boolean)
Invert contrast of rescaled image
percentile: (0.0 <= a floating point number <= 50.0, nipype default
value: 0.0)
Percentile to use for reference to allow for outliers - 1 indicates
the 1st and 99th percentiles in the input file will be mapped to the
99th and 1st percentiles in the reference; 0 indicates minima and
maxima will be mapped
Outputs:
out_file: (an existing file name)
Rescaled image