TypeError: 'SwigPyObject' object is not subscriptable
```
### 1.2. Software
```
importitkimportsysprint("Self-compiled version of ITK:\t", itk. __version__ )print("Python version:\t", sys.version)
```
```
Self-compiled version of ITK: 5.3.0
Python version: 3.10.4 (main, Mar 23 2022, 23:05:40) [GCC 11.2.0]
```
---
### Author:  [@matt.mccormick](https://discourse.itk.org/u/matt.mccormick)
#### Post date: [May 13, 2022, 4:56pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/2 "2022-05-13T16:56:45Z")
Hi @edgar ,
To resample an image with `itk.resample_image_filter` and a b-spline transform, pass the transform in as the `transform` argument.
For the input, use the moving image. The sampling grid can be specified with a reference image, e.g. the fixed image, or by the parameters of the sampling gride, including the ‘size’.
HTH,
Matt
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [May 16, 2022, 11:33am UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/3 "2022-05-16T11:33:37Z")
Hi and thanks, @matt.mccormick. Sorry for the wall of text. In [`DeformableRegistration15.cxx`](https://itk.org/Doxygen/html/Examples_2RegistrationITKv4_2DeformableRegistration15_8cxx-example.html), `GetCoefficientImages()[k]` is used within a loop to provide the input image. How can I access that image in Python? I get an error if I try (see above). Thanks!
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [May 23, 2022, 5:21pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/4 "2022-05-23T17:21:26Z")
Hi, I’m sure everyone is busy, and it’s alright. If someone has the time, I would appreciate the help. Thanks.
---
### Author:  [@Pranjal\_Sahu](https://discourse.itk.org/u/Pranjal_Sahu)
#### Post date: [May 27, 2022, 3:35pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/5 "2022-05-27T15:35:29Z")
> [@edgar](#):
>
> ```auto
> Wrong number or type of arguments for overloaded function 'itkImageToImageFilterIF3IF3_SetInput'.
> Possible C/C++ prototypes are:
> itkImageToImageFilterIF3IF3::SetInput(itkImageF3 const *)
> itkImageToImageFilterIF3IF3::SetInput(unsigned int,itkImageF3 const *)
>
> ```
Can you please check if the data type is correct ?
It expects float format as per the error message.
---
### Author:  [@Pranjal\_Sahu](https://discourse.itk.org/u/Pranjal_Sahu)
#### Post date: [May 27, 2022, 3:40pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/6 "2022-05-27T15:40:01Z")
> [@edgar](#):
>
> `itkImageToImageFilterIF3IF3`
Also please post a minimum end-to-end Python code to reproduce the error.
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [May 29, 2022, 8:19pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/7 "2022-05-29T20:19:51Z")
About the code: sure, here it is.
```
#!/usr/bin/env python
# This code is partially based on
# DeformableRegistration15.cxx
# which holds the Apache 2.0 license (see below).
#
# Copyright 2022 edgar
# License:
#
# This program is free software: you can redistribute it
# and/or modify it under the terms of the GNU General
# Public License as published by the Free Software
# Foundation, version 3 of the License.
#
# This program is distributed in the hope that it will be
# useful, but WITHOUT ANY WARRANTY; without even the
# implied warranty of MERCHANTABILITY or FITNESS FOR A
# PARTICULAR PURPOSE. See the GNU General Public License
# for more details.
#
# You should have received a copy of the GNU General
# Public License along with this program. If not, see
# .
#
# Copyright Insight Software Consortium
#
# Licensed under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License. You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0.txt
#
# Unless required by applicable law or agreed to in writing,
# software distributed under the License is distributed on
# an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
# KIND, either express or implied. See the License for the
# specific language governing permissions and limitations
# under the License.
import sys
import itk
import numpy as np
from distutils.version import StrictVersion as VS
from scipy import ndimage as ndi
from skimage.morphology import disk
from scipy.ndimage import shift
if VS(itk.Version.GetITKVersion()) < VS("5.3.0"):
print("ITK 5.3.0 is required.")
sys.exit(1)
# Draw a 3D disk
side_size = 10
obj2d = disk(side_size / 2, dtype=np.uint8)
obj3d = np.asarray([disk(side_size / 2, dtype=np.uint8)
for idx in range(5)])
# Create 4 frames
series_length = 4
holder = np.asarray([obj3d for idx in range(series_length)])
# Move the last frame
holder[-1] = shift(holder[-1],
(series_length - 1, 0, 0))
fixed_image_arr = holder[0].transpose((2, 1, 0))
moving_image_arr = holder[-1].transpose((2, 1, 0))
fixed_image = itk.image_from_array(
fixed_image_arr.astype(np.float32))
moving_image = itk.image_from_array(
moving_image_arr.astype(np.float32))
spatial_dim = fixed_image.GetImageDimension()
fixed_image_type = type(fixed_image)
moving_image_type = type(moving_image)
origin = [0] * spatial_dim
length_per_px_ratios = [1] * spatial_dim
[dict(img).update({
# Set origin ([0, 0, 0]); fixed_image.GetOrigin()
"origin": origin,
# Axial scale (length unit / pixel for each axis)
"spacing": length_per_px_ratios,
# Set orientation of axes; fixed_image.GetDirection()
"direction": np.eye(spatial_dim)})
for img in (fixed_image, moving_image)]
coord_type = itk.D
versor_type = coord_type
versor_transform_class = itk.VersorRigid3DTransform[versor_type]
rigid_transform = versor_transform_class.New()
trnfm_init_class = itk.CenteredTransformInitializer[
# versor_transform_class,
itk.VersorRigid3DTransform[itk.D],
fixed_image_type,
moving_image_type]
optimiser_class = itk.RegularStepGradientDescentOptimizer
optimiser = optimiser_class.New()
metric = itk.MattesMutualInformationImageToImageMetric[
fixed_image_type, moving_image_type].New(
NumberOfHistogramBins=50)
metric.ReinitializeSeed(76926294)
metric.SetUseCachingOfBSplineWeights(True)
metric.SetUseExplicitPDFDerivatives(True)
interpolator = itk.LinearInterpolateImageFunction[
moving_image_type, coord_type].New()
registrant = itk.ImageRegistrationMethod[
fixed_image_type, moving_image_type].New(
Metric=metric,
Optimizer=optimiser,
Interpolator=interpolator,
FixedImage=fixed_image,
MovingImage=moving_image)
identity_transform = itk.IdentityTransform[
coord_type, spatial_dim].New()
initialiser = trnfm_init_class.New(
Transform=rigid_transform,
FixedImage=fixed_image,
MovingImage=moving_image)
initialiser.MomentsOn()
initialiser.InitializeTransform()
fixed_region = fixed_image.GetBufferedRegion()
number_of_pixels = fixed_region.GetNumberOfPixels()
registrant.SetFixedImageRegion(fixed_region)
registrant.SetInitialTransformParameters(
rigid_transform.GetParameters())
registrant.SetTransform(rigid_transform)
translation_scale = 1 / 1000
optimiser_scales = [1.0, 1.0, 1.0,
translation_scale,
translation_scale,
translation_scale]
optimiser.SetScales(optimiser_scales)
optimiser.SetMaximumStepLength(0.2000)
optimiser.SetMinimumStepLength(0.0001)
optimiser.SetNumberOfIterations(200)
metric.SetNumberOfSpatialSamples(int(10000))
registrant.Update()
rigid_transform.SetParameters(registrant.GetLastTransformParameters())
affine_transform_class = itk.AffineTransform[
coord_type, spatial_dim]
affine_transform = affine_transform_class.New(
Center=rigid_transform.GetCenter(),
Translation=rigid_transform.GetTranslation(),
Matrix=rigid_transform.GetMatrix())
registrant.SetTransform(affine_transform)
registrant.SetInitialTransformParameters(
affine_transform.GetParameters())
optimiser.SetScales([1] * 9 + [translation_scale] * 3)
optimiser.SetMaximumStepLength(0.2)
optimiser.SetMinimumStepLength(1e-4)
optimiser.SetNumberOfIterations(200)
metric.SetNumberOfSpatialSamples(int(50000))
registrant.Update()
affine_transform.SetParameters(registrant.GetLastTransformParameters())
spline_order = 3
deformable_transform_class = itk.BSplineTransform[
coord_type, spatial_dim, spline_order]
bspline_transform_coarse = deformable_transform_class.New()
grid_nodes_1_dim_coarse = int(5)
fixed_origin = fixed_image.GetOrigin()
fixed_physical_dimensions = [fixed_image.GetSpacing()[idx] * (
fixed_image.GetLargestPossibleRegion().GetSize()[idx] - 1)
for idx in range(spatial_dim)]
bspline_transform_coarse.SetTransformDomainOrigin(fixed_origin)
bspline_transform_coarse.SetTransformDomainPhysicalDimensions(
fixed_physical_dimensions)
mesh_size = \
bspline_transform_coarse.GetTransformDomainMeshSize()
mesh_size.Fill(grid_nodes_1_dim_coarse - spline_order)
bspline_transform_coarse.SetTransformDomainMeshSize(mesh_size)
bspline_transform_coarse.SetTransformDomainDirection(
fixed_image.GetDirection())
number_of_bspline_param = \
bspline_transform_coarse.GetNumberOfParameters()
optimiser.SetScales(
[1.0 for idx in range(number_of_bspline_param)])
composite_transform_class = itk.CompositeTransform[
coord_type, spatial_dim]
composite_transform = composite_transform_class.New()
composite_transform.AddTransform(affine_transform)
composite_transform.AddTransform(bspline_transform_coarse)
composite_transform.SetOnlyMostRecentTransformToOptimizeOn()
bspline_transform_coarse.GetParameters().Fill(1.0)
registrant.SetInitialTransformParameters(
bspline_transform_coarse.GetParameters())
registrant.SetTransform(composite_transform)
optimiser.SetMaximumStepLength(10.0)
optimiser.SetMinimumStepLength(0.01)
optimiser.SetRelaxationFactor(0.7)
optimiser.SetNumberOfIterations(50)
metric.SetNumberOfSpatialSamples(number_of_bspline_param * 100)
registrant.Update()
final_parameters = registrant.GetLastTransformParameters()
bspline_transform_coarse.SetParameters(final_parameters)
bspline_transform_fine = deformable_transform_class.New(
TransformDomainOrigin=fixed_origin,
TransformDomainPhysicalDimensions=fixed_physical_dimensions,
TransformDomainMeshSize=mesh_size,
TransformDomainDirection=fixed_image.GetDirection())
grid_nodes_1_dim_fine = int(20)
mesh_size.Fill(grid_nodes_1_dim_fine - spline_order)
number_of_bspline_param = \
bspline_transform_fine.GetNumberOfParameters()
interpolator_fine = itk.BSplineResampleImageFunction[
moving_image_type, coord_type].New()
upsampler = itk.ResampleImageFilter[moving_image_type, moving_image_type].New(
Input=bspline_transform_coarse.GetCoefficientImages(),
Interpolator=interpolator_fine,
Transform=identity_transform,
Size=bspline_transform_fine.GetCoefficientImages())
upsampler = itk.resample_image_filter[moving_image_type, moving_image_type](
Input=bspline_transform_coarse.GetCoefficientImages()[0],
interpolator=interpolator_fine,
transform=identity_transform,
size=bspline_transform_fine.GetCoefficientImages())
```
About the data, the idea is to use `GetCoefficientImages` as the input (just as in the example), but let me know if this is the information that you need:
```
moving_image_type:
fixed_image_type:
```
Thank you!
---
### Author:  [@Pranjal\_Sahu](https://discourse.itk.org/u/Pranjal_Sahu)
#### Post date: [June 2, 2022, 1:28pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/8 "2022-06-02T13:28:34Z")
If I print the object using the following code,
```auto
t1 = bspline_transform_coarse.GetCoefficientImages()
print(t1)
```
I get:
```auto
>,3 > *' at 0x7f0e70fa7480>
```
which is a fixed array of image with double type. So you can’t directly pass this as input.
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [June 2, 2022, 7:28pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/9 "2022-06-02T19:28:35Z")
Thanks. In [https://itk.org/Doxygen/html/Examples\_2RegistrationITKv4\_2DeformableRegistration15\_8cxx-example.html](https://itk.org/Doxygen/html/Examples_2RegistrationITKv4_2DeformableRegistration15_8cxx-example.html), they use
`bspline_transform_coarse.GetCoefficientImages()[k]`. Is there a way to do the same with Python? How to access the k^{th} element? Is it even possible?
---
### Author:  [@Pranjal\_Sahu](https://discourse.itk.org/u/Pranjal_Sahu)
#### Post date: [June 2, 2022, 8:09pm UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/10 "2022-06-02T20:09:16Z")
> [@edgar](#):
>
> [ITK: Examples/RegistrationITKv4/DeformableRegistration15.cxx](https://itk.org/Doxygen/html/Examples_2RegistrationITKv4_2DeformableRegistration15_8cxx-example.html)
You can first get the deformation field and then upsample it and then decompose it to BSpline.
```auto
convertFilter = itk.TransformToDisplacementFieldFilter.IVF33D.New()
convertFilter.SetTransform(bsplineTransform)
convertFilter.UseReferenceImageOn()
convertFilter.SetReferenceImage(fixedImage)
convertFilter.Update()
field = convertFilter.GetOutput()
field = np.array(field)
np.save('displacement_field.npy', field)
```
Here, upsample the field and then pass it as input to the [itk::BSplineDecompositionImageFilter](https://itk.org/Doxygen/html/classitk_1_1BSplineDecompositionImageFilter.html) to create a new B-spline transform for next iteration.
I will have to check if there is a way to image from GetCoefficientImages method in Python.
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [June 3, 2022, 7:52am UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/11 "2022-06-03T07:52:22Z")
> [@Pranjal\_Sahu](#):
>
> I will have to check if there is a way to image from GetCoefficientImages method in Python.
Ok, thanks.
---
### Author:  [@edgar](https://discourse.itk.org/u/edgar)
#### Post date: [August 16, 2022, 9:06am UTC](https://discourse.itk.org/t/how-to-itk-resample-image-filter-with-a-bsplinetransform-with-3d-images/4987/12 "2022-08-16T09:06:57Z")
Will there be an implementation?