# Volume Cropping in Python Script

**URL:** <https://discourse.slicer.org/t/volume-cropping-in-python-script/2585>\
**Category:** Support\
**Created:** [April 13, 2018, 8:09pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585 "2018-04-13T20:09:56Z")\
**Posts on this page:** 9\
**Page:** 1

<div class="post-metadata">

**Author:** ![brhoom](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/brhoom/32/1228_2.png) [@brhoom](https://discourse.slicer.org/u/brhoom)\
**Post date:** [April 13, 2018, 8:09pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/1 "2018-04-13T20:09:56Z")

</div>

Dear all,  
I just checked [Python\_scripting](https://www.slicer.org/wiki/Documentation/Nightly/Developers/Python_scripting) and [ScriptRepository](https://www.slicer.org/wiki/Documentation/Nightly/ScriptRepository) pages and couldn’t find an example for cropping an image, I wrote this simple example. It runs without problem inside a module or extension but when I run it in an external python script it also runs and produces the correct output cropped image but there are error messages I don’t understand. Here is the mentioned complete script code:

```
 from __main__ import vtk, qt, ctk, slicer
 import sitkUtils

 print(" myScript ")

 inputVolumeFnm = ./myVolume.nrrd"  
 croppedImageFnm = "./myCroppedVolume.nrrd"

 [success, inputVolume] = slicer.util.loadVolume(inputVolumeFnm, returnNode=True)          
 inputImage = sitkUtils.PullVolumeFromSlicer(inputVolume.GetID())

 cropper = sitkUtils.sitk.CropImageFilter()
 croppingBounds = [[178, 210, 67],[227, 195, 34]]  
 croppedImage = cropper.Execute(inputImage, croppingBounds[0], croppingBounds[1]) 
 croppedNode = sitkUtils.PushVolumeToSlicer(croppedImage, None, inputVolume.GetName() , 
 'vtkMRMLScalarVolumeNode' )

 properties = {}
 properties["fileType"] = ".nrrd"
 slicer.util.saveNode(croppedNode, croppedImageFnm, properties)

 print(" All tasks are done! ") 

 exit()

```

Despite the correct output cropped image, I get this output in the terminal:

```
~/sw/Slicer-4.8.1/Slicer --no-main-window --python-script ./myScript.py
Qt: Cannot set locale modifiers: 
Number of registered modules: 150 
Number of instantiated modules: 150 
Number of loaded modules: 150 
               myScript 
Loaded volume from file: ./myVolume.nrrd. Dimensions: 485x485x121. Number of components: 1. Pixel type: short.
"Volume" Reader has successfully read the file "./myVolume.nrrd" "[0.56s]" 
vtkITKArchetypeImageSeriesReader::ExecuteInformation: Archetype file ./_DUMMY_DOES_NOT_EXIST__ does not exist.
Algorithm vtkITKArchetypeDiffusionTensorImageReaderFile(0x4a977b0) returned failure for request: 
vtkInformation (0x4a9aea0)
 Debug: Off
 Modified Time: 827149
 Reference Count: 1
 Registered Events: (none)
 Request: REQUEST_INFORMATION
 ALGORITHM_AFTER_FORWARD: 1
 FORWARD_DIRECTION: 0
vtkITKArchetypeImageSeriesReader::ExecuteInformation: Archetype file ./_DUMMY_DOES_NOT_EXIST__ does not exist.
 Algorithm vtkITKArchetypeImageSeriesVectorReaderSeries(0x48edfb0) returned failure for request: vtkInformation (0x4a978f0)
  Debug: Off
  Modified Time: 828705
  Reference Count: 1
  Registered Events: (none)
 Request: REQUEST_INFORMATION
  ALGORITHM_AFTER_FORWARD: 1
  FORWARD_DIRECTION: 0
 vtkITKArchetypeImageSeriesReader::ExecuteInformation: Archetype file /_DUMMY_DOES_NOT_EXIST__ 
 does not exist.
 Algorithm vtkITKArchetypeImageSeriesScalarReader(0x4a6be60) returned failure for request: vtkInformation (0x4a8c600)
  Debug: Off
  Modified Time: 829510
  Reference Count: 1
  Registered Events: (none)
  Request: REQUEST_INFORMATION
  ALGORITHM_AFTER_FORWARD: 1
  FORWARD_DIRECTION: 0
  ReadData: This is not a nrrd file
  ReadData: Cannot read file as a volume of type DiffusionTensorVolume[fullName 
         /_DUMMY_DOES_NOT_EXIST__]
Number of files listed in the node = 0.
File reader says it was able to read 0 files.
File reader used the archetype file name of //_DUMMY_DOES_NOT_EXIST__ []
  FileNotFoundError
  ReadData: This is not a nrrd file
  ReadData: Failed to instantiate a file reader
  ReadData: Cannot read file as a volume of type Volume[fullName = /_DUMMY_DOES_NOT_EXIST__]
Number of files listed in the node = 0.
File reader says it was able to read 0 files.
File reader used the archetype file name of /_DUMMY_DOES_NOT_EXIST__ []
   FileNotFoundError
  Generic Warning: In /home/kitware/Dashboards/Package/Slicer- 
 481/Libs/MRML/Core/vtkDataFileFormatHelper.cxx, line 237
 vtkDataFileFormatHelper::GetFileExtensionFromFormatString: please update deprecated extension-only 
 format specifier to 'File format name (.ext)' format! Current format string: .nrrd
 All tasks are done!
```

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [April 14, 2018, 5:10am UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/2 "2018-04-14T05:10:06Z")

</div>

This warning is generated by a [workaround in sitkUtils](https://github.com/Slicer/Slicer/blob/69358596167cbc99d39acfa98a5cfa2a1e4042dd/Base/Python/sitkUtils.py#L50-L63). You can safely ignore the warnings.

@blowekamp Do you think we still need this workaround? On Windows, it does not seem to be necessary - for example, this code works:

```
import SimpleITK as sitk

# Pull volume from Slicer
inputVolumeNode = getNode('MRHead')
inputVolumeNodeSceneAddress = inputVolumeNode.GetScene().GetAddressAsString("").replace('Addr=','')
inputVolumeNodeFullITKAddress = 'slicer:{0}#{1}'.format(inputVolumeNodeSceneAddress,inputVolumeNode.GetID())
sitkimage = sitk.ReadImage(inputVolumeNodeFullITKAddress)

# Process
filter = sitk.SignedMaurerDistanceMapImageFilter()
outputImage = filter.Execute(sitkimage)

# Push volume to Slicer
outputVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode")
outputVolumeNode.CreateDefaultDisplayNodes()
outputVolumeNodeSceneAddress = outputVolumeNode.GetScene().GetAddressAsString("").replace('Addr=','')
outputVolumeNodeFullITKAddress = 'slicer:{0}#{1}'.format(inputVolumeNodeSceneAddress,outputVolumeNode.GetID())
sitk.WriteImage(outputImage, outputVolumeNodeFullITKAddress)
slicer.util.setSliceViewerLayers(background = outputVolumeNode)
```

---

<div class="post-metadata">

**Author:** ![blowekamp](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/blowekamp/32/1386_2.png) [@blowekamp](https://discourse.slicer.org/u/blowekamp)\
**Post date:** [April 16, 2018, 1:34pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/3 "2018-04-16T13:34:22Z")

</div>

That code appears to be added by yourself:

> <https://github.com/Slicer/Slicer/commit/1128611e088884b8030e4b278a3c4976aaf04239>

The ITK factory registration is a complicated process and likely has changed a bit since I last looked at it in Slicer so I don’t know.

With SimpleITK 1.1 we have added a method to get a list of the registered ITK ImageIO. This could detect if this need to happen or not.  
[https://itk.org/SimpleITKDoxygen110/html/classitk\_1\_1simple\_1\_1ImageReaderBase.html#a6d5066467465ebff4b2bb895e7ed59d0](https://itk.org/SimpleITKDoxygen110/html/classitk_1_1simple_1_1ImageReaderBase.html#a6d5066467465ebff4b2bb895e7ed59d0)

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [April 16, 2018, 2:09pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/4 "2018-04-16T14:09:58Z")

</div>

I’ve tried to access [std::vector\<std::string\> itk::simple::ImageReaderBase::GetRegisteredImageIOs()](https://itk.org/SimpleITKDoxygen110/html/classitk_1_1simple_1_1ImageReaderBase.html#a6d5066467465ebff4b2bb895e7ed59d0) from Python but it did not seem to be available. Do you know if `std::vector<std::string>` types are supposed to be wrapped for Python in SimpleITK? (I know that VTK wrappers cannot deal with such types)

I just reorganized the code, the code for ensuring registration was added by @hjmjohnson.

---

<div class="post-metadata">

**Author:** ![blowekamp](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/blowekamp/32/1386_2.png) [@blowekamp](https://discourse.slicer.org/u/blowekamp)\
**Post date:** [April 16, 2018, 2:23pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/5 "2018-04-16T14:23:47Z")

</div>

Currently, Slicer is using 1.0.1-ish there is a [pull request to update to 1.1.0.](https://github.com/Slicer/Slicer/pull/929).

If you get SimpleITK with pip it will be there.

Yes the return type is [wrapped](https://github.com/SimpleITK/SimpleITK/blob/master/Wrapping/Common/SimpleITK_Common.i#L99) and [tested](https://github.com/SimpleITK/SimpleITK/blob/554c7304e108ad6c51122b28f23a57c0e722b110/Testing/Unit/sitkImageIOTests.cxx) in SimpleITK.

```auto

In [1]: import SimpleITK as sitk

In [2]: print sitk.Version()
SimpleITK Version: 1.1.0 (ITK 4.13)
Compiled: Mar 22 2018 21:14:44

In [3]: print sitk.ImageFileReader().GetRegisteredImageIOs()
('BMPImageIO', 'BioRadImageIO', 'Bruker2dseqImageIO', 'GDCMImageIO', 'GE4ImageIO', 'GE5ImageIO', 'GiplImageIO', 'HDF5ImageIO', 'JPEGImageIO', 'LSMImageIO', 'MINCImageIO', 'MRCImageIO', 'MetaImageIO', 'NiftiImageIO', 'NrrdImageIO', 'PNGImageIO', 'StimulateImageIO', 'TIFFImageIO', 'VTKImageIO')

```

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [April 16, 2018, 8:41pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/6 "2018-04-16T20:41:05Z")

</div>

Great! I’ll test Slicer with SimpleITK 1.1.0 tomorrow and if everything works then I’ll update the nightly build to use that version.

---

<div class="post-metadata">

**Author:** ![zjx1805](https://avatars.discourse-cdn.com/v4/letter/z/f08c70/32.png) [@zjx1805](https://discourse.slicer.org/u/zjx1805)\
**Post date:** [April 17, 2018, 2:56am UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/7 "2018-04-17T02:56:39Z")

</div>

Hello @brhoom, I saw your post here and thought you might be familiar with cropping volume in Python. I am wondering if you know of a way to crop a volume using ROI in 3D slicer in Python, modify the cropped volume and put it back into the original volume (at the same position). Essentially, what I am trying to do is like this

```
Drag some ROI in 3D slicer
croppedVolumeNode = some_internal_cropping_function(volumeNode, roiNode)
modify the cropped volume
volumeNode.update(croppedVolumeNode) # to put it back to the original location

```

Thanks!

---

<div class="post-metadata">

**Author:** ![brhoom](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/brhoom/32/1228_2.png) [@brhoom](https://discourse.slicer.org/u/brhoom)\
**Post date:** [April 22, 2018, 6:02pm UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/8 "2018-04-22T18:02:22Z")

</div>

> [@zjx1805](#):
>
> I saw your post here and thought you might be familiar with cropping volume in Python

Sorry for the late reply. I do the cropping using only one point, the center of the cropping. So to get the croppingBounds in my code above from a defined point:

```
    spacing = inputVolume.GetSpacing()
    imgData = inputVolume.GetImageData()
    dimensions = imgData.GetDimensions() 
    size = [lengthX,lengthY,lengthZ] # defined volume dimensions    
    # Calculate lower and upper cropping bounds
    lower = [0,0,0]
    for i in range(0,3):
        lower[i] = point[i] - int((size[i]/spacing[i])/2)
    upper = [0,0,0]
    for i in range(0,3):
        upper[i] = dimensions[i] - (point[i]+int((size[i]/spacing[i])/2))
    croppingBounds = [lower,upper]

```

I think in your case, I will do the following 🙂

- create a copy of the original volume.
- do the cropping
- process the cropped volume.
- convert the copied volume and the cropped volume to np arrays and replace the voxels in the first volume with the cropped one.

Probably there is a better way. I am not the Slicer expert here. Also, I think it is not a good idea to ask different question in different post. I would open a new post with the new question to keep the forum organized.

Hope this helps.

---

<div class="post-metadata">

**Author:** ![zjx1805](https://avatars.discourse-cdn.com/v4/letter/z/f08c70/32.png) [@zjx1805](https://discourse.slicer.org/u/zjx1805)\
**Post date:** [April 23, 2018, 5:08am UTC](https://discourse.slicer.org/t/volume-cropping-in-python-script/2585/9 "2018-04-23T05:08:46Z")

</div>

Yeah I posted a separate question in the forum. Thank you for your reply anyway!