Script repository

Note

Usage: Copy-paste the code lines displayed below or the linked .py file contents into Python console in Slicer. Or save them to a .py file and run them using execfile.

Launch Slicer

Open a file with Slicer at the command line

Open imagefile.nrrd file in Slicer:

Slicer.exe /full/path/to/imagefile.nrrd

Note

It may be necessary to specify full path to the Slicer executable and to the file that needs to be loaded.

To load a file with non-default options, you can use --python-code option to run slicer.util.load... commands.

Open an .mrb file with Slicer at the command line

Slicer.exe --python-code "slicer.util.loadScene('f:/2013-08-23-Scene.mrb')"

Run Python commands in the Slicer environment

Run Python commands, without showing any graphical user interface:

Slicer.exe --python-code "doSomething; doSomethingElse; etc." --testing --no-splash --no-main-window

Slicer exits when the commands are completed because --testing options is specified.

Run a Python script file in the Slicer environment

Run a Python script (stored in script file), without showing any graphical user interface:

Slicer.exe --python-script "/full/path/to/myscript.py" --no-splash --no-main-window

To make Slicer exit when the script execution is completed, call sys.exit(errorCode) (where errorCode is set 0 for success and other value to indicate error).

MRML scene

Get MRML node from the scene

Get markups fiducial node named F (useful for quickly getting access to a MRML node in the Python console):

fidsNode = getNode('F')
# do something with the node... let's remove the first control point in it
fidsNode.RemoveNthControlPoint(0)

Getting the first volume node without knowing its name (useful if there is only one volume loaded):

volumeNode = slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLScalarVolumeNode")
# do something with the node... let's change its display window/level
volumeNode.GetDisplayNode().SetAutoWindowLevel(False)
volumeNode.GetDisplayNode().SetWindowLevelMinMax(100, 200)

Note

  • slicer.util.getNode() is recommended only for interactive debugging in the Python console/Jupyter notebook
    • its input is intentionally defined vaguely (it can be either node ID or name and you can use wildcards such as *), which is good because it make it simpler to use, but the uncertain behavior is not good for general-purpose use in a module
    • throws an exception so that the developer knows immediately that there was a typo or other unexpected error
  • slicer.mrmlScene.GetNodeByID() is more appropriate when a module needs to access a MRML node:
    • its behavior is more predictable: it only accepts node ID as input. slicer.mrmlScene.GetFirstNodeByName() can be used to get a node by its name, but since multiple nodes in the scene can have the same name, it is not recommended to keep reference to a node by its name. Since node IDs may change when a scene is saved and reloaded, node ID should not be stored persistently, but node references must be used instead
    • if node is not found it returns None (instead of throwing an exception), because this is often not considered an error in module code (it is just used to check existence of a node) and using return value for not-found nodes allows simpler syntax

Clone a node

This example shows how to make a copy of any node that appears in Subject Hierarchy (in Data module).

# Get a node from SampleData that we will clone
import SampleData
nodeToClone = SampleData.SampleDataLogic().downloadMRHead()

# Clone the node
shNode = slicer.vtkMRMLSubjectHierarchyNode.GetSubjectHierarchyNode(slicer.mrmlScene)
itemIDToClone = shNode.GetItemByDataNode(nodeToClone)
clonedItemID = slicer.modules.subjecthierarchy.logic().CloneSubjectHierarchyItem(shNode, itemIDToClone)
clonedNode = shNode.GetItemDataNode(clonedItemID)

Save a node to file

Save a transform node to file (should work with any other node type, if file extension is set to a supported one):

myNode = getNode("LinearTransform_3")

myStorageNode = myNode.CreateDefaultStorageNode()
myStorageNode.SetFileName("c:/tmp/something.tfm")
myStorageNode.WriteData(myNode)

Save the scene into a single MRB file

# Generate file name
import time
sceneSaveFilename = slicer.app.temporaryPath + "/saved-scene-" + time.strftime("%Y%m%d-%H%M%S") + ".mrb"

# Save scene
if slicer.util.saveScene(sceneSaveFilename):
  logging.info("Scene saved to: {0}".format(sceneSaveFilename))
else:
  logging.error("Scene saving failed")

Save the scene into a new directory

# Create a new directory where the scene will be saved into
import time
sceneSaveDirectory = slicer.app.temporaryPath + "/saved-scene-" + time.strftime("%Y%m%d-%H%M%S")
if not os.access(sceneSaveDirectory, os.F_OK):
  os.makedirs(sceneSaveDirectory)

# Save the scene
if slicer.app.applicationLogic().SaveSceneToSlicerDataBundleDirectory(sceneSaveDirectory, None):
  logging.info("Scene saved to: {0}".format(sceneSaveDirectory))
else:
  logging.error("Scene saving failed")

Override default scene save dialog

Place this class in the scripted module file to override

class MyModuleFileDialog ():
   """This specially named class is detected by the scripted loadable
   module and is the target for optional drag and drop operations.
   See: Base/QTGUI/qSlicerScriptedFileDialog.h.

   This class is used for overriding default scene save dialog
   with simple saving the scene without asking anything.
   """

   def __init__(self,qSlicerFileDialog ):
     self.qSlicerFileDialog = qSlicerFileDialog
     qSlicerFileDialog.fileType = "NoFile"
     qSlicerFileDialog.description = "Save scene"
     qSlicerFileDialog.action = slicer.qSlicerFileDialog.Write

   def execDialog(self):
     # Implement custom scene save operation here.
     # Return True if saving completed successfully,
     # return False if saving was cancelled.
     ...
     return saved

Override application close behavior

When application close is requested then by default confirmation popup is displayed. To customize this behavior (for example, allow application closing without displaying default confirmation popup) an event filter can be installed for the close event on the main window:

class CloseApplicationEventFilter(qt.QWidget):
  def eventFilter(self, object, event):
    if event.type() == qt.QEvent.Close:
      event.accept()
      return True
    return False

filter = CloseApplicationEventFilter()
slicer.util.mainWindow().installEventFilter(filter)

Change default output file type for new nodes

This script changes default output file format for nodes that have not been saved yet (do not have storage node yet).

Default node can be specified that will be used as a basis of all new storage nodes. This can be used for setting default file extension. For example, change file format to PLY for model nodes:

defaultModelStorageNode = slicer.vtkMRMLModelStorageNode()
defaultModelStorageNode.SetDefaultWriteFileExtension("ply")
slicer.mrmlScene.AddDefaultNode(defaultModelStorageNode)

To permanently change default file extension on your computer, copy-paste the code above into your application startup script (you can find its location in menu: Edit / Application settings / General / Application startup script).

Change file type for saving for existing nodes

This script changes output file types for nodes that have been already saved (they already have storage node).

If it is not necessary to preserve file paths then the simplest is to configure default storage node (as shown in the example above), then delete all existing storage nodes. When save dialog is opened, default storage nodes will be recreated.

# Delete existing model storage nodes so that they will be recreated with default settings
existingModelStorageNodes = slicer.util.getNodesByClass("vtkMRMLModelStorageNode")
for modelStorageNode in existingModelStorageNodes:
  slicer.mrmlScene.RemoveNode(modelStorageNode)

To update existing storage nodes to use new file extension (but keep all other parameters unchanged) you can use this approach (example is for volume storage):

requiredFileExtension = ".nia"
originalFileExtension = ".nrrd"
volumeNodes = slicer.util.getNodesByClass("vtkMRMLScalarVolumeNode")
for volumeNode in volumeNodes:
  volumeStorageNode = volumeNode.GetStorageNode()
  if not volumeStorageNode:
    volumeNode.AddDefaultStorageNode()
    volumeStorageNode = volumeNode.GetStorageNode()
    volumeStorageNode.SetFileName(volumeNode.GetName()+requiredFileExtension)
  else:
    volumeStorageNode.SetFileName(volumeStorageNode.GetFileName().replace(originalFileExtension, requiredFileExtension))

To set all volume nodes to save uncompressed by default (add this to .slicerrc.py file so it takes effect for the whole session):

#set the default volume storage to not compress by default
defaultVolumeStorageNode = slicer.vtkMRMLVolumeArchetypeStorageNode()
defaultVolumeStorageNode.SetUseCompression(0)
slicer.mrmlScene.AddDefaultNode(defaultVolumeStorageNode)
logging.info("Volume nodes will be stored uncompressed by default")

Same thing as above, but applied to all segmentations instead of volumes:

#set the default volume storage to not compress by default
defaultVolumeStorageNode = slicer.vtkMRMLSegmentationStorageNode()
defaultVolumeStorageNode.SetUseCompression(0)
slicer.mrmlScene.AddDefaultNode(defaultVolumeStorageNode)
logging.info("Segmentation nodes will be stored uncompressed

Module selection

Switch to a different module

This utility function can be used to open a different module:

slicer.util.selectModule("DICOM")

Set a new default module at startup

Instead of the default Welcome module:

qt.QSettings().setValue("Modules/HomeModule", "Data")

Views

Display text in a 3D view or slice view

The easiest way to show information overlaid on a viewer is to use corner annotations.

view=slicer.app.layoutManager().threeDWidget(0).threeDView()
# Set text to "Something"
view.cornerAnnotation().SetText(vtk.vtkCornerAnnotation.UpperRight,"Something")
# Set color to red
view.cornerAnnotation().GetTextProperty().SetColor(1,0,0)
# Update the view
view.forceRender()

To display text in slice views, replace the first line by this line (and consider hiding slice view annotations, to prevent them from overwriting the text you place there):

view=slicer.app.layoutManager().sliceWidget("Red").sliceView()

Show orientation marker in all views

viewNodes = slicer.util.getNodesByClass("vtkMRMLAbstractViewNode")
for viewNode in viewNodes:
  viewNode.SetOrientationMarkerType(slicer.vtkMRMLAbstractViewNode.OrientationMarkerTypeAxes)

Change view axis labels

labels = ["x", "X", "y", "Y", "z", "Z"]
viewNode = slicer.app.layoutManager().threeDWidget(0).mrmlViewNode()
# for slice view:
# viewNode = slicer.app.layoutManager().sliceWidget("Red").mrmlSliceNode()
for index, label in enumerate(labels):
  viewNode.SetAxisLabel(index, label)

Hide view controller bars

slicer.app.layoutManager().threeDWidget(0).threeDController().setVisible(False)
slicer.app.layoutManager().sliceWidget("Red").sliceController().setVisible(False)
slicer.app.layoutManager().plotWidget(0).plotController().setVisible(False)
slicer.app.layoutManager().tableWidget(0).tableController().setVisible(False)

Hide Slicer logo from main window

This script increases vertical space available in the module panel by hiding the Slicer application logo.

slicer.util.findChild(slicer.util.mainWindow(), "LogoLabel").visible = False

Customize widgets in view controller bars

sliceController = slicer.app.layoutManager().sliceWidget("Red").sliceController()

# hide what is not needed
sliceController.pinButton().hide()
#sliceController.viewLabel().hide()
sliceController.fitToWindowToolButton().hide()
sliceController.sliceOffsetSlider().hide()

# add custom widgets
myButton = qt.QPushButton("My custom button")
sliceController.barLayout().addWidget(myButton)

Get current mouse coordinates in a slice view

You can get 3D (RAS) coordinates of the current mouse cursor from the crosshair singleton node as shown in the example below:

def onMouseMoved(observer,eventid):
  ras=[0,0,0]
  crosshairNode.GetCursorPositionRAS(ras)
  print(ras)

crosshairNode=slicer.util.getNode("Crosshair")
crosshairNode.AddObserver(slicer.vtkMRMLCrosshairNode.CursorPositionModifiedEvent, onMouseMoved)

Display crosshair at a 3D position

position_RAS = [23.4, 5.6, 78.9]
crosshairNode = slicer.util.getNode("Crosshair")
# Set crosshair position
crosshairNode.SetCrosshairRAS(position_RAS)
# Center the position in all slice views
slicer.vtkMRMLSliceNode.JumpAllSlices(slicer.mrmlScene, *position_RAS, slicer.vtkMRMLSliceNode.CenteredJumpSlice)

Note

Crosshair node stores two positions: Cursor position is the current position of the mouse pointer in a slice or 3D view (modules should only read this position). Crosshair position is the location of the visible crosshair in views (modules can read or write this position).

Display mouse pointer coordinates in alternative coordinate system

The Data probe only shows coordinate values in the world coordinate system. You can make the world coordinate system mean anything you want (e.g., MNI) by applying a transform to the volume that transforms it into that space. See more details in here.

def onMouseMoved(observer,eventid):
  mniToWorldTransformNode = getNode("LinearTransform_3")  # replace this by the name of your actual MNI to world transform
  worldToMniTransform = vtk.vtkGeneralTransform()
  mniToWorldTransformNode.GetTransformToWorld(worldToMniTransform)
  ras=[0,0,0]
  mni=[0,0,0]
  crosshairNode.GetCursorPositionRAS(ras)
  worldToMniTransform.TransformPoint(ras, mni)
  _ras = "; ".join([str(k) for k in ras])
  _mni = "; ".join([str(k) for k in mni])
  slicer.util.showStatusMessage(f"RAS={_ras}   MNI={_mni}")

crosshairNode=slicer.util.getNode("Crosshair")
observationId = crosshairNode.AddObserver(slicer.vtkMRMLCrosshairNode.CursorPositionModifiedEvent, onMouseMoved)

# Run this to stop displaying values:
# crosshairNode.RemoveObserver(observationId)

Get DataProbe text

You can get the mouse location in pixel coordinates along with the pixel value at the mouse by hitting the . (period) key in a slice view after pasting in the following code.

def printDataProbe():
  infoWidget = slicer.modules.DataProbeInstance.infoWidget
  for layer in ("B", "F", "L"):
    print(infoWidget.layerNames[layer].text, infoWidget.layerIJKs[layer].text, infoWidget.layerValues[layer].text)

s = qt.QShortcut(qt.QKeySequence("."), mainWindow())
s.connect("activated()", printDataProbe)

Create custom color table

This example shows how to create a new color table, for example with inverted color range from the default Ocean color table.

invertedocean = slicer.vtkMRMLColorTableNode()
invertedocean.SetTypeToUser()
invertedocean.SetNumberOfColors(256)
invertedocean.SetName("InvertedOcean")

for i in range(0,255):
  invertedocean.SetColor(i, 0.0, 1 - (i+1e-16)/255.0, 1.0, 1.0)

slicer.mrmlScene.AddNode(invertedocean)

Show color scalar bar in slice views

Display color bar for background volume in slice views (managed by DataProbe):

sliceAnnotations = slicer.modules.DataProbeInstance.infoWidget.sliceAnnotations
sliceAnnotations.sliceViewAnnotationsEnabled = True
sliceAnnotations.scalarBarEnabled = 1
sliceAnnotations.scalarBarSelectedLayer = "background"  # alternative is "foreground"
sliceAnnotations.rangeLabelFormat = "test %G"
sliceAnnotations.updateSliceViewFromGUI()

Display color scalar bar in 3D views

colorTableRangeMm = 40
title ="Radial\nCompression\n"
labelsFormat = "%4.1f mm"

# Create color node
colorNode = slicer.mrmlScene.CreateNodeByClass("vtkMRMLProceduralColorNode")
colorNode.UnRegister(None)  # to prevent memory leaks
colorNode.SetName(slicer.mrmlScene.GenerateUniqueName("MyColormap"))
colorNode.SetAttribute("Category", "MyModule")
# The color node is a procedural color node, which is saved using a storage node.
# Hidden nodes are not saved if they use a storage node, therefore
# the color node must be visible.
colorNode.SetHideFromEditors(False)
slicer.mrmlScene.AddNode(colorNode)

# Specify colormap
colorMap = colorNode.GetColorTransferFunction()
colorMap.RemoveAllPoints()
colorMap.AddRGBPoint(colorTableRangeMm * 0.0, 0.0, 0.0, 1.0)
colorMap.AddRGBPoint(colorTableRangeMm * 0.2, 0.0, 1.0, 1.0)
colorMap.AddRGBPoint(colorTableRangeMm * 0.5, 1.0, 1.0, 0.0)
colorMap.AddRGBPoint(colorTableRangeMm * 1.0, 1.0, 0.0, 0.0)

# Display color scalar bar
colorWidget = slicer.modules.colors.widgetRepresentation()
colorWidget.setCurrentColorNode(colorNode)
ctkScalarBarWidget = slicer.util.findChildren(colorWidget, name="VTKScalarBar")[0]
ctkScalarBarWidget.setDisplay(1)
ctkScalarBarWidget.setTitle(title)
ctkScalarBarWidget.setMaxNumberOfColors(256)
ctkScalarBarWidget.setLabelsFormat(labelsFormat)

Customize view layout

Show a custom layout of a 3D view on top of the red slice view:

customLayout = """
<layout type="vertical" split="true">
  <item>
   <view class="vtkMRMLViewNode" singletontag="1">
     <property name="viewlabel" action="default">1</property>
   </view>
  </item>
  <item>
   <view class="vtkMRMLSliceNode" singletontag="Red">
    <property name="orientation" action="default">Axial</property>
    <property name="viewlabel" action="default">R</property>
    <property name="viewcolor" action="default">#F34A33</property>
   </view>
  </item>
</layout>
"""

# Built-in layout IDs are all below 100, so you can choose any large random number
# for your custom layout ID.
customLayoutId=501

layoutManager = slicer.app.layoutManager()
layoutManager.layoutLogic().GetLayoutNode().AddLayoutDescription(customLayoutId, customLayout)

# Switch to the new custom layout
layoutManager.setLayout(customLayoutId)

See description of standard layouts (that can be used as examples) here: https://github.com/Slicer/Slicer/blob/master/Libs/MRML/Logic/vtkMRMLLayoutLogic.cxx

You can use this code snippet to add a button to the layout selector toolbar:

# Add button to layout selector toolbar for this custom layout
viewToolBar = mainWindow().findChild("QToolBar", "ViewToolBar")
layoutMenu = viewToolBar.widgetForAction(viewToolBar.actions()[0]).menu()
layoutSwitchActionParent = layoutMenu  # use `layoutMenu` to add inside layout list, use `viewToolBar` to add next the standard layout list
layoutSwitchAction = layoutSwitchActionParent.addAction("My view") # add inside layout list
layoutSwitchAction.setData(layoutId)
layoutSwitchAction.setIcon(qt.QIcon(":Icons/Go.png"))
layoutSwitchAction.setToolTip("3D and slice view")

Turn on slice intersections

viewNodes = slicer.util.getNodesByClass("vtkMRMLSliceCompositeNode")
for viewNode in viewNodes:
  viewNode.SetSliceIntersectionVisibility(1)

Note

How to find code corresponding to a user interface widget?

For this one I searched for “slice intersections” text in the whole slicer source code, found that the function is implemented in Base\QTGUI\qSlicerViewersToolBar.cxx, then translated the qSlicerViewersToolBarPrivate::setSliceIntersectionVisible(bool visible) method to Python.

Hide slice view annotations

This script can hide node name, patient information displayed in corners of slice views (managed by DataProbe module).

# Disable slice annotations immediately
sliceAnnotations = slicer.modules.DataProbeInstance.infoWidget.sliceAnnotations
sliceAnnotations.sliceViewAnnotationsEnabled=False
sliceAnnotations.updateSliceViewFromGUI()
# Disable slice annotations persistently (after Slicer restarts)
settings = qt.QSettings()
settings.setValue("DataProbe/sliceViewAnnotations.enabled", 0)

Change slice offset

Equivalent to moving the slider in slice view controller.

layoutManager = slicer.app.layoutManager()
red = layoutManager.sliceWidget("Red")
redLogic = red.sliceLogic()
# Print current slice offset position
print(redLogic.GetSliceOffset())
# Change slice position
redLogic.SetSliceOffset(20)

Change slice orientation

Get Red slice node and rotate around X and Y axes.

sliceNode = slicer.app.layoutManager().sliceWidget("Red").mrmlSliceNode()
sliceToRas = sliceNode.GetSliceToRAS()
transform=vtk.vtkTransform()
transform.SetMatrix(SliceToRAS)
transform.RotateX(20)
transform.RotateY(15)
sliceToRas.DeepCopy(transform.GetMatrix())
sliceNode.UpdateMatrices()

Measure angle between two slice planes

Measure angle between red and yellow slice nodes. Whenever any of the slice nodes are moved, the updated angle is printed on the console.

sliceNodeIds = ["vtkMRMLSliceNodeRed", "vtkMRMLSliceNodeYellow"]

# Print angles between slice nodes
def ShowAngle(unused1=None, unused2=None):
  sliceNormalVector = []
  for sliceNodeId in sliceNodeIds:
    sliceToRAS = slicer.mrmlScene.GetNodeByID(sliceNodeId).GetSliceToRAS()
    sliceNormalVector.append([sliceToRAS.GetElement(0,2), sliceToRAS.GetElement(1,2), sliceToRAS.GetElement(2,2)])
  angleRad = vtk.vtkMath.AngleBetweenVectors(sliceNormalVector[0], sliceNormalVector[1])
  angleDeg = vtk.vtkMath.DegreesFromRadians(angleRad)
  print("Angle between slice planes = {0:0.3f}".format(angleDeg))

# Observe slice node changes
for sliceNodeId in sliceNodeIds:
  slicer.mrmlScene.GetNodeByID(sliceNodeId).AddObserver(vtk.vtkCommand.ModifiedEvent, ShowAngle)

# Print current angle
ShowAngle()

Set slice position and orientation from a normal vector and position

This code snippet shows how to display a slice view defined by a normal vector and position in an anatomically sensible way: rotating slice view so that “up” direction (or “right” direction) is towards an anatomical axis.

def setSlicePoseFromSliceNormalAndPosition(sliceNode, sliceNormal, slicePosition, defaultViewUpDirection=None, backupViewRightDirection=None):
  """
  Set slice pose from the provided plane normal and position. View up direction is determined automatically,
  to make view up point towards defaultViewUpDirection.
  :param defaultViewUpDirection Slice view will be spinned in-plane to match point approximately this up direction. Default: patient superior.
  :param backupViewRightDirection Slice view will be spinned in-plane to match point approximately this right direction
    if defaultViewUpDirection is too similar to sliceNormal. Default: patient left.
  """
  # Fix up input directions
  if defaultViewUpDirection is None:
    defaultViewUpDirection = [0,0,1]
  if backupViewRightDirection is None:
    backupViewRightDirection = [-1,0,0]
  if sliceNormal[1]>=0:
    sliceNormalStandardized = sliceNormal
  else:
    sliceNormalStandardized = [-sliceNormal[0], -sliceNormal[1], -sliceNormal[2]]
  # Compute slice axes
  sliceNormalViewUpAngle = vtk.vtkMath.AngleBetweenVectors(sliceNormalStandardized, defaultViewUpDirection)
  angleTooSmallThresholdRad = 0.25 # about 15 degrees
  if sliceNormalViewUpAngle > angleTooSmallThresholdRad and sliceNormalViewUpAngle < vtk.vtkMath.Pi() - angleTooSmallThresholdRad:
    viewUpDirection = defaultViewUpDirection
    sliceAxisY = viewUpDirection
    sliceAxisX = [0, 0, 0]
    vtk.vtkMath.Cross(sliceAxisY, sliceNormalStandardized, sliceAxisX)
  else:
    sliceAxisX = backupViewRightDirection
  # Set slice axes
  sliceNode.SetSliceToRASByNTP(sliceNormalStandardized[0], sliceNormalStandardized[1], sliceNormalStandardized[2],
    sliceAxisX[0], sliceAxisX[1], sliceAxisX[2],
    slicePosition[0], slicePosition[1], slicePosition[2], 0)

# Example usage:
sliceNode = getNode("vtkMRMLSliceNodeRed")
transformNode = getNode("Transform_3")
transformMatrix = vtk.vtkMatrix4x4()
transformNode.GetMatrixTransformToParent(transformMatrix)
sliceNormal = [transformMatrix.GetElement(0,2), transformMatrix.GetElement(1,2), transformMatrix.GetElement(2,2)]
slicePosition = [transformMatrix.GetElement(0,3), transformMatrix.GetElement(1,3), transformMatrix.GetElement(2,3)]
setSlicePoseFromSliceNormalAndPosition(sliceNode, sliceNormal, slicePosition)

Show slice views in 3D window

Equivalent to clicking ‘eye’ icon in the slice view controller.

layoutManager = slicer.app.layoutManager()
for sliceViewName in layoutManager.sliceViewNames():
  controller = layoutManager.sliceWidget(sliceViewName).sliceController()
  controller.setSliceVisible(True)

Change default slice view orientation

You can left-right “flip” slice view orientation presets (show patient left side on left/right side of the screen) by copy-pasting the script below to your .slicerrc.py file.

# Axial slice axes:
#  1 0 0
#  0 1 0
#  0 0 1
axialSliceToRas=vtk.vtkMatrix3x3()

# Coronal slice axes:
#  1 0 0
#  0 0 -1
#  0 1 0
coronalSliceToRas=vtk.vtkMatrix3x3()
coronalSliceToRas.SetElement(1,1, 0)
coronalSliceToRas.SetElement(1,2, -1)
coronalSliceToRas.SetElement(2,1, 1)
coronalSliceToRas.SetElement(2,2, 0)

# Replace orientation presets in all existing slice nodes and in the default slice node
sliceNodes = slicer.util.getNodesByClass("vtkMRMLSliceNode")
sliceNodes.append(slicer.mrmlScene.GetDefaultNodeByClass("vtkMRMLSliceNode"))
for sliceNode in sliceNodes:
  orientationPresetName = sliceNode.GetOrientation()
  sliceNode.RemoveSliceOrientationPreset("Axial")
  sliceNode.AddSliceOrientationPreset("Axial", axialSliceToRas)
  sliceNode.RemoveSliceOrientationPreset("Coronal")
  sliceNode.AddSliceOrientationPreset("Coronal", coronalSliceToRas)
  sliceNode.SetOrientation(orientationPresetName)

Set all slice views linked by default

You can make slice views linked by default (when application starts or the scene is cleared) by copy-pasting the script below to your .slicerrc.py file.

# Set linked slice views  in all existing slice composite nodes and in the default node
sliceCompositeNodes = slicer.util.getNodesByClass("vtkMRMLSliceCompositeNode")
defaultSliceCompositeNode = slicer.mrmlScene.GetDefaultNodeByClass("vtkMRMLSliceCompositeNode")
if not defaultSliceCompositeNode:
  defaultSliceCompositeNode = slicer.mrmlScene.CreateNodeByClass("vtkMRMLSliceCompositeNode")
  defaultSliceCompositeNode.UnRegister(None)  # CreateNodeByClass is factory method, need to unregister the result to prevent memory leaks
  slicer.mrmlScene.AddDefaultNode(defaultSliceCompositeNode)
sliceCompositeNodes.append(defaultSliceCompositeNode)
for sliceCompositeNode in sliceCompositeNodes:
  sliceCompositeNode.SetLinkedControl(True)

Set crosshair jump mode to centered by default

You can change default slice jump mode (when application starts or the scene is cleared) by copy-pasting the script below to your .slicerrc.py file.

crosshair=slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLCrosshairNode")
crosshair.SetCrosshairBehavior(crosshair.CenteredJumpSlice)

Set up custom units in slice view ruler

For microscopy or micro-CT images you may want to switch unit to micrometer instead of the default mm. To do that, 1. change the unit in Application settings / Units and 2. update ruler display settings using the script below (it can be copied to your Application startup script):

lm = slicer.app.layoutManager()
for sliceViewName in lm.sliceViewNames():
  sliceView = lm.sliceWidget(sliceViewName).sliceView()
  displayableManager = sliceView.displayableManagerByClassName("vtkMRMLRulerDisplayableManager")
  displayableManager.RemoveAllRulerScalePresets()
  displayableManager.AddRulerScalePreset(   0.001, 5, 2, "nm", 1000.0)
  displayableManager.AddRulerScalePreset(   0.010, 5, 2, "nm", 1000.0)
  displayableManager.AddRulerScalePreset(   0.100, 5, 2, "nm", 1000.0)
  displayableManager.AddRulerScalePreset(   0.500, 5, 1, "nm", 1000.0)
  displayableManager.AddRulerScalePreset(   1.0,   5, 2, "um",    1.0)
  displayableManager.AddRulerScalePreset(   5.0,   5, 1, "um",    1.0)
  displayableManager.AddRulerScalePreset(  10.0,   5, 2, "um",    1.0)
  displayableManager.AddRulerScalePreset(  50.0,   5, 1, "um",    1.0)
  displayableManager.AddRulerScalePreset( 100.0,   5, 2, "um",    1.0)
  displayableManager.AddRulerScalePreset( 500.0,   5, 1, "um",    1.0)
  displayableManager.AddRulerScalePreset(1000.0,   5, 2, "mm",    0.001)

Center the 3D view on the scene

layoutManager = slicer.app.layoutManager()
threeDWidget = layoutManager.threeDWidget(0)
threeDView = threeDWidget.threeDView()
threeDView.resetFocalPoint()

Rotate the 3D View

layoutManager = slicer.app.layoutManager()
threeDWidget = layoutManager.threeDWidget(0)
threeDView = threeDWidget.threeDView()
threeDView.yaw()

Change 3D view background color

viewNode = slicer.app.layoutManager().threeDWidget(0).mrmlViewNode()
viewNode.SetBackgroundColor(1,0,0)
viewNode.SetBackgroundColor2(1,0,0)

Show a slice view outside the view layout

# layout name is used to create and identify the underlying slice node and  should be set to a value that is not used in any of the layouts owned by the layout manager
layoutName = "TestSlice1"
layoutLabel = "TS1"
layoutColor = [1.0, 1.0, 0.0]
# ownerNode manages this view instead of the layout manager (it can be any node in the scene)
viewOwnerNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScriptedModuleNode")

# Create MRML nodes
viewLogic = slicer.vtkMRMLSliceLogic()
viewLogic.SetMRMLScene(slicer.mrmlScene)
viewNode = viewLogic.AddSliceNode(layoutName)
viewNode.SetLayoutLabel(layoutLabel)
viewNode.SetLayoutColor(layoutColor)
viewNode.SetAndObserveParentLayoutNodeID(viewOwnerNode.GetID())

# Create widget
viewWidget = slicer.qMRMLSliceWidget()
viewWidget.setMRMLScene(slicer.mrmlScene)
viewWidget.setMRMLSliceNode(viewNode)
sliceLogics = slicer.app.applicationLogic().GetSliceLogics()
viewWidget.setSliceLogics(sliceLogics)
sliceLogics.AddItem(viewWidget.sliceLogic())
viewWidget.show()

Show a 3D view outside the view layout

# layout name is used to create and identify the underlying view node and  should be set to a value that is not used in any of the layouts owned by the layout manager
layoutName = "Test3DView"
layoutLabel = "T3"
layoutColor = [1.0, 1.0, 0.0]
# ownerNode manages this view instead of the layout manager (it can be any node in the scene)
viewOwnerNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScriptedModuleNode")

# Create MRML node
viewLogic = slicer.vtkMRMLViewLogic()
viewLogic.SetMRMLScene(slicer.mrmlScene)
viewNode = viewLogic.AddViewNode(layoutName)
viewNode.SetLayoutLabel(layoutLabel)
viewNode.SetLayoutColor(layoutColor)
viewNode.SetAndObserveParentLayoutNodeID(viewOwnerNode.GetID())

# Create widget
viewWidget = slicer.qMRMLThreeDWidget()
viewWidget.setMRMLScene(slicer.mrmlScene)
viewWidget.setMRMLViewNode(viewNode)
viewWidget.show()

Access VTK rendering classes

Accesss VTK views, renderers, and cameras

Iterate through all 3D views in current layout:

layoutManager = slicer.app.layoutManager()
for threeDViewIndex in range(layoutManager.threeDViewCount) :
  view = layoutManager.threeDWidget(threeDViewIndex).threeDView()
  threeDViewNode = view.mrmlViewNode()
  cameraNode = slicer.modules.cameras.logic().GetViewActiveCameraNode(threeDViewNode)
  print("View node for 3D widget " + str(threeDViewIndex))
  print("  Name: " + threeDViewNode .GetName())
  print("  ID: " + threeDViewNode .GetID())
  print("  Camera ID: " + cameraNode.GetID())

Iterate through all slice views in current layout:

layoutManager = slicer.app.layoutManager()
for sliceViewName in layoutManager.sliceViewNames():
  view = layoutManager.sliceWidget(sliceViewName).sliceView()
  sliceNode = view.mrmlSliceNode()
  sliceLogic = slicer.app.applicationLogic().GetSliceLogic(sliceNode)
  compositeNode = sliceLogic.GetSliceCompositeNode()
  print("Slice view " + str(sliceViewName))
  print("  Name: " + sliceNode.GetName())
  print("  ID: " + sliceNode.GetID())
  print("  Background volume: {0}".format(compositeNode.GetBackgroundVolumeID()))
  print("  Foreground volume: {0} (opacity: {1})".format(compositeNode.GetForegroundVolumeID(), compositeNode.GetForegroundOpacity()))
  print("  Label volume: {0} (opacity: {1})".format(compositeNode.GetLabelVolumeID(), compositeNode.GetLabelOpacity()))

For low-level manipulation of views, it is possible to access VTK render windows, renderers and cameras of views in the current layout.

renderWindow = view.renderWindow()
renderers = renderWindow.GetRenderers()
renderer = renderers.GetItemAsObject(0)
camera = cameraNode.GetCamera()

Get displayable manager of a certain type for a certain view

Displayable managers are responsible for creating VTK filters, mappers, and actors to display MRML nodes in renderers. Input to filters and mappers are VTK objects stored in MRML data nodes. Filter and actor properties are set based on display options specified in MRML display nodes.

Accessing displayable managers is useful for troubleshooting or for testing new features that are not exposed via MRML classes yet, as they provide usually allow low-level access to VTK actors.

threeDViewWidget = slicer.app.layoutManager().threeDWidget(0)
modelDisplayableManager = threeDViewWidget.threeDView().displayableManagerByClassName("vtkMRMLModelDisplayableManager")
if modelDisplayableManager is None:
  logging.error("Failed to find the model displayable manager")

Access VTK actor properties

This example shows how to access and modify VTK actor properties to experiment with physically-based rendering.

modelNode = slicer.util.getNode("MyModel")

threeDViewWidget = slicer.app.layoutManager().threeDWidget(0)
modelDisplayableManager = threeDViewWidget.threeDView().displayableManagerByClassName("vtkMRMLModelDisplayableManager")
actor=modelDisplayableManager.GetActorByID(modelNode.GetDisplayNode().GetID())
property=actor.GetProperty()
property.SetInterpolationToPBR()
property.SetMetallic(0.5)
property.SetRoughness(0.5)
property.SetColor(0.5,0.5,0.9)
slicer.util.forceRenderAllViews()

See more information on physically based rendering in VTK here: https://blog.kitware.com/vtk-pbr/

Keyboard shortcuts and mouse gestures

Customize keyboard shortcuts

Keyboard shortcuts can be specified for activating any Slicer feature by adding a couple of lines to your .slicerrc.py file.

For example, this script registers Ctrl+b, Ctrl+n, Ctrl+m, Ctrl+, keyboard shortcuts to switch between red, yellow, green, and 4-up view layouts.

shortcuts = [
    ("Ctrl+b", lambda: slicer.app.layoutManager().setLayout(slicer.vtkMRMLLayoutNode.SlicerLayoutOneUpRedSliceView)),
    ("Ctrl+n", lambda: slicer.app.layoutManager().setLayout(slicer.vtkMRMLLayoutNode.SlicerLayoutOneUpYellowSliceView)),
    ("Ctrl+m", lambda: slicer.app.layoutManager().setLayout(slicer.vtkMRMLLayoutNode.SlicerLayoutOneUpGreenSliceView)),
    ("Ctrl+,", lambda: slicer.app.layoutManager().setLayout(slicer.vtkMRMLLayoutNode.SlicerLayoutFourUpView))
    ]

for (shortcutKey, callback) in shortcuts:
    shortcut = qt.QShortcut(slicer.util.mainWindow())
    shortcut.setKey(qt.QKeySequence(shortcutKey))
    shortcut.connect( "activated()", callback)

Here’s an example for cycling through Segment Editor effects (requested on the forum for the SlicerMorph project).

def cycleEffect(delta=1):
  try:
    orderedNames = list(slicer.modules.SegmentEditorWidget.editor.effectNameOrder())
    allNames = slicer.modules.SegmentEditorWidget.editor.availableEffectNames()
    for name in allNames:
      try:
        orderedNames.index(name)
      except ValueError:
        orderedNames.append(name)
    orderedNames.insert(0, None)
    activeEffect = slicer.modules.SegmentEditorWidget.editor.activeEffect()
    if activeEffect:
      activeName = slicer.modules.SegmentEditorWidget.editor.activeEffect().name
    else:
      activeName = None
    newIndex = (orderedNames.index(activeName) + delta) % len(orderedNames)
    slicer.modules.SegmentEditorWidget.editor.setActiveEffectByName(orderedNames[newIndex])
  except AttributeError:
    # module not active
    pass

shortcuts = [
  ("`", lambda: cycleEffect(-1)),
  ("~", lambda: cycleEffect(1)),
  ]

for (shortcutKey, callback) in shortcuts:
  shortcut = qt.QShortcut(slicer.util.mainWindow())
  shortcut.setKey(qt.QKeySequence(shortcutKey))
  shortcut.connect( "activated()", callback)

Customize keyboard/mouse gestures in viewers

Example for making the 3D view rotate using right-click-and-drag:

threeDViewWidget = slicer.app.layoutManager().threeDWidget(0)
cameraDisplayableManager = threeDViewWidget.threeDView().displayableManagerByClassName("vtkMRMLCameraDisplayableManager")
cameraWidget = cameraDisplayableManager.GetCameraWidget()

# Remove old mapping from right-click-and-drag
cameraWidget.SetEventTranslationClickAndDrag(cameraWidget.WidgetStateIdle, vtk.vtkCommand.RightButtonPressEvent, vtk.vtkEvent.NoModifier,
  cameraWidget.WidgetStateRotate, vtk.vtkWidgetEvent.NoEvent, vtk.vtkWidgetEvent.NoEvent)

# Make right-click-and-drag rotate the view
cameraWidget.SetEventTranslationClickAndDrag(cameraWidget.WidgetStateIdle, vtk.vtkCommand.RightButtonPressEvent, vtk.vtkEvent.NoModifier,
  cameraWidget.WidgetStateRotate, cameraWidget.WidgetEventRotateStart, cameraWidget.WidgetEventRotateEnd)

Disable certain user interactions in slice views

For example, disable slice browsing using mouse wheel and keyboard shortcuts in the red slice viewer:

interactorStyle = slicer.app.layoutManager().sliceWidget("Red").sliceView().sliceViewInteractorStyle()
interactorStyle.SetActionEnabled(interactorStyle.BrowseSlice, False)

Hide all slice view controllers:

lm = slicer.app.layoutManager()
for sliceViewName in lm.sliceViewNames():
  lm.sliceWidget(sliceViewName).sliceController().setVisible(False)

Hide all 3D view controllers:

lm = slicer.app.layoutManager()
for viewIndex in range(slicer.app.layoutManager().threeDViewCount):
  lm.threeDWidget(0).threeDController().setVisible(False)

Add keyboard shortcut to jump to center or world coordinate system

You can copy-paste this into the Python console to jump slice views to (0,0,0) position on (Ctrl+e):

shortcut = qt.QShortcut(qt.QKeySequence("Ctrl+e"), slicer.util.mainWindow())
shortcut.connect("activated()",
  lambda: slicer.modules.markups.logic().JumpSlicesToLocation(0,0,0, True))

Launch external applications

How to run external applications from Slicer.

Launch external process in startup environment

When a process is launched from Slicer then by default Slicer”s ITK, VTK, Qt, etc. libraries are used. If an external application has its own version of these libraries, then the application is expected to crash. To prevent crashing, the application must be run in the environment where Slicer started up (without all Slicer-specific library paths). This startup environment can be retrieved using slicer.util.startupEnvironment().

Example: run Python3 script from Slicer:

command_to_execute = ["/usr/bin/python3", "-c", "print("hola")"]
from subprocess import check_output
check_output(
  command_to_execute,
  env=slicer.util.startupEnvironment()
  )

will output:

"hola\n"

On some systems, shell=True must be specified as well.

Manage extensions

Download and install extension

extensionName = 'SlicerIGT'
em = slicer.app.extensionsManagerModel()
if not em.isExtensionInstalled(extensionName):
  extensionMetaData = em.retrieveExtensionMetadataByName(extensionName)
  url = em.serverUrl().toString()+'/download/item/'+extensionMetaData['item_id']
  extensionPackageFilename = slicer.app.temporaryPath+'/'+extensionMetaData['md5']
  slicer.util.downloadFile(url, extensionPackageFilename)
  em.installExtension(extensionPackageFilename)
  slicer.util.restart()

Install a module directly from a git repository

This code snippet can be useful for sharing code in development without requiring a restart of Slicer.

DICOM

Load DICOM files into the scene from a folder

This code loads all DICOM objects into the scene from a file folder. All the registered plugins are evaluated and the one with the highest confidence will be used to load the data. Files are imported into a temporary DICOM database, so the current Slicer DICOM database is not impacted.

dicomDataDir = "c:/my/folder/with/dicom-files"  # input folder with DICOM files
loadedNodeIDs = []  # this list will contain the list of all loaded node IDs

from DICOMLib import DICOMUtils
with DICOMUtils.TemporaryDICOMDatabase() as db:
  DICOMUtils.importDicom(dicomDataDir, db)
  patientUIDs = db.patients()
  for patientUID in patientUIDs:
    loadedNodeIDs.extend(DICOMUtils.loadPatientByUID(patientUID))

Import DICOM files into the application’s DICOM database

This code snippet uses Slicer DICOM browser built-in indexer to import DICOM files into the database. Images are not loaded into the scene, but they show up in the DICOM browser. After import, data sets can be loaded using DICOMUtils functions (e.g., loadPatientByUID) - see above for an example.

# instantiate a new DICOM browser
slicer.util.selectModule("DICOM")
dicomBrowser = slicer.modules.DICOMWidget.browserWidget.dicomBrowser
# use dicomBrowser.ImportDirectoryCopy to make a copy of the files (useful for importing data from removable storage)
dicomBrowser.importDirectory(dicomFilesDirectory, dicomBrowser.ImportDirectoryAddLink)
# wait for import to finish before proceeding (optional, if removed then import runs in the background)
dicomBrowser.waitForImportFinished()

Import DICOM files using DICOMweb

Download and import DICOM data set using DICOMweb from Kheops, Google Health API, etc.

How to obtain accessToken:

  • Google Cloud: Execute gcloud auth print-access-token once you have logged in
  • Kheops: create an album, create a sharing link (somethin like `https://demo.kheops.online/view/TfYXwbKAW7JYbAgZ7MyISf <https://demo.kheops.online/view/TfYXwbKAW7JYbAgZ7MyISf>`__), the token is the string after the last slash
slicer.util.selectModule("DICOM")  # ensure DICOM database is initialized and
slicer.app.processEvents()
from DICOMLib import DICOMUtils
DICOMUtils.importFromDICOMWeb(
    dicomWebEndpoint="http://demo.kheops.online/api",
    studyInstanceUID="1.3.6.1.4.1.14519.5.2.1.8421.4009.985792766370191766692237040819",
    accessToken="TfYXwbKAW7JYbAgZ7MyISf")

Access top level tags of DICOM images imported into Slicer

For example, to print the first patient’s first study’s first series’ “0020,0032” field:

db = slicer.dicomDatabase
patientList = db.patients()
studyList = db.studiesForPatient(patientList[0])
seriesList = db.seriesForStudy(studyList[0])
fileList = db.filesForSeries(seriesList[0])
# Note, fileValue accesses the database of cached top level tags
# (nested tags are not included)
print(db.fileValue(fileList[0], "0020,0032"))
# Get tag group,number from dicom dictionary
import pydicom as dicom
tagName = "StudyDate"
tagStr = str(dicom.tag.Tag(tagName))[1:-1].replace(" ","")
print(db.fileValue(fileList[0], tagStr))

Access DICOM tags nested in a sequence

db = slicer.dicomDatabase
patientList = db.patients()
studyList = db.studiesForPatient(patientList[0])
seriesList = db.seriesForStudy(studyList[0])
fileList = db.filesForSeries(seriesList[0])
# Use pydicom to access the full header, which requires
# re-reading the dataset instead of using the database cache
import pydicom
pydicom.dcmread(fileList[0])
ds.CTExposureSequence[0].ExposureModulationType

Access tag of a volume loaded from DICOM? For example, get the patient position stored in a volume

volumeName = "2: ENT IMRT"
n = slicer.util.getNode(volumeName)
instUids = n.GetAttribute("DICOM.instanceUIDs").split()
filename = slicer.dicomDatabase.fileForInstance(instUids[0])
print(slicer.dicomDatabase.fileValue(filename, "0018,5100"))

Access tag of an item in the Subject Hierachy tree

For example, get the content time tag of a structure set:

rtStructName = "3: RTSTRUCT: PROS"
rtStructNode = slicer.util.getNode(rtStructName)
shNode = slicer.vtkMRMLSubjectHierarchyNode.GetSubjectHierarchyNode(slicer.mrmlScene)
rtStructShItemID = shNode.GetItemByDataNode(rtStructNode)
ctSliceInstanceUids = shNode.GetItemAttribute(rtStructShItemID, "DICOM.ReferencedInstanceUIDs").split()
filename = slicer.dicomDatabase.fileForInstance(ctSliceInstanceUids[0])
print(slicer.dicomDatabase.fileValue(filename, "0008,0033"))

Get path and filename of a loaded DICOM volume

def pathFromNode(node):
  storageNode = node.GetStorageNode()
  if storageNode is not None: # loaded via drag-drop
      filepath = storageNode.GetFullNameFromFileName()
  else: # Loaded via DICOM browser
      instanceUIDs = node.GetAttribute("DICOM.instanceUIDs").split()
      filepath = slicer.dicomDatabase.fileForInstance(instUids[0])
  return filepath

# Example:
node = slicer.util.getNode("volume1")
path = self.pathFromNode(node)
print("DICOM path=%s" % path)

Convert DICOM to NRRD on the command line

/Applications/Slicer-4.6.2.app/Contents/MacOS/Slicer --no-main-window --python-code "node=slicer.util.loadVolume('/tmp/series/im0.dcm'); slicer.util.saveNode(node, "/tmp/output.nrrd"); exit()"

The same can be done on windows by using the top level Slicer.exe. Be sure to use forward slashes in the pathnames within quotes on the command line.

Export a volume to DICOM file format

volumeNode = getNode("CTChest")
outputFolder = "c:/tmp/dicom-output"

# Create patient and study and put the volume under the study
shNode = slicer.vtkMRMLSubjectHierarchyNode.GetSubjectHierarchyNode(slicer.mrmlScene)
patientItemID = shNode.CreateSubjectItem(shNode.GetSceneItemID(), "test patient")
studyItemID = shNode.CreateStudyItem(patientItemID, "test study")
volumeShItemID = shNode.GetItemByDataNode(volumeNode)
shNode.SetItemParent(volumeShItemID, studyItemID)

import DICOMScalarVolumePlugin
exporter = DICOMScalarVolumePlugin.DICOMScalarVolumePluginClass()
exportables = exporter.examineForExport(volumeShItemID)
for exp in exportables:
  exp.directory = outputFolder

exporter.export(exportables)

Export a segmentation to DICOM segmentation object

segmentationNode = ...
referenceVolumeNode = ...
outputFolder = "c:/tmp/dicom-output"

# Associate segmentation node with a reference volume node
shNode = slicer.vtkMRMLSubjectHierarchyNode.GetSubjectHierarchyNode(slicer.mrmlScene)
referenceVolumeShItem = shNode.GetItemByDataNode(referenceVolumeNode)
studyShItem = shNode.GetItemParent(referenceVolumeShItem)
segmentationShItem = shNode.GetItemByDataNode(segmentationNode)
shNode.SetItemParent(segmentationShItem, studyShItem)

# Export to DICOM
import DICOMSegmentationPlugin
exporter = DICOMSegmentationPlugin.DICOMSegmentationPluginClass()
exportables = exporter.examineForExport(segmentationShItem)
for exp in exportables:
  exp.directory = outputFolder

exporter.export(exportables)

Customize table columns in DICOM browser

Documentation of methods for changing DICOM browser columns: https://github.com/commontk/CTK/blob/master/Libs/DICOM/Core/ctkDICOMDatabase.h#L354-L375

# Get browser and database
dicomBrowser = slicer.modules.dicom.widgetRepresentation().self().browserWidget.dicomBrowser
dicomDatabase = dicomBrowser.database()

# Print list of available columns
print(dicomDatabase.patientFieldNames)
print(dicomDatabase.studyFieldNames)
print(dicomDatabase.seriesFieldNames)

# Change column order
dicomDatabase.setWeightForField("Series", "SeriesDescription", 7)
dicomDatabase.setWeightForField("Studies", "StudyDescription", 6)
# Change column visibility
dicomDatabase.setVisibilityForField("Patients", "PatientsBirthDate", False)
dicomDatabase.setVisibilityForField("Patients", "PatientsComments", True)
dicomDatabase.setWeightForField("Patients", "PatientsComments", 8)
# Change column name
dicomDatabase.setDisplayedNameForField("Series", "DisplayedCount", "Number of images")
# Change column width to manual
dicomDatabase.setFormatForField("Series", "SeriesDescription", '{"resizeMode":"interactive"}')
# Customize table manager in DICOM browser
dicomTableManager = dicomBrowser.dicomTableManager()
dicomTableManager.selectionMode = qt.QAbstractItemView.SingleSelection
dicomTableManager.autoSelectSeries = False

# Force database views update
dicomDatabase.closeDatabase()
dicomDatabase.openDatabase(dicomBrowser.database().databaseFilename)

Query and retrieve data from a PACS using classic DIMSE DICOM networking

# Query
dicomQuery = ctk.ctkDICOMQuery()
dicomQuery.callingAETitle = "SLICER"
dicomQuery.calledAETitle = "ANYAE"
dicomQuery.host = "dicomserver.co.uk"
dicomQuery.port = 11112
dicomQuery.preferCGET = True
dicomQuery.filters = {"Name":"Anon", "Modalities":"MR"}
# temporary in-memory database for storing query results
tempDb = ctk.ctkDICOMDatabase()
tempDb.openDatabase("")
dicomQuery.query(tempDb)

# Retrieve
dicomRetrieve = ctk.ctkDICOMRetrieve()
dicomRetrieve.callingAETitle = dicomQuery.callingAETitle
dicomRetrieve.calledAETitle = dicomQuery.calledAETitle
dicomRetrieve.host = dicomQuery.host
dicomRetrieve.port = dicomQuery.port
dicomRetrieve.setMoveDestinationAETitle("SLICER");
dicomRetrieve.setDatabase(slicer.dicomDatabase)
for study in dicomQuery.studyInstanceUIDQueried:
    print(f"ctkDICOMRetrieveTest2: Retrieving {study}")
    slicer.app.processEvents()
    if dicomQuery.preferCGET:
        success = dicomRetrieve.getStudy(study)
    else:
        success = dicomRetrieve.moveStudy(study)
    print(f"  - {'success' if success else 'failed'}")
slicer.dicomDatabase.updateDisplayedFields()

Convert RT structure set to labelmap NRRD files

SlicerRT batch processing to batch convert RT structure sets to labelmap NRRD files.

Markups

Load markups fiducial list from file

Markups fiducials can be loaded from file:

slicer.util.loadMarkupsFiducialList("/path/to/list/F.fcsv")

Adding Fiducials Programatically

Markups fiducials can be added to the currently active list from the python console by using the following module logic command:

slicer.modules.markups.logic().AddFiducial()

The command with no arguments will place a new fiducial at the origin. You can also pass it an initial location:

slicer.modules.markups.logic().AddFiducial(1.0, -2.0, 3.3)

How to draw a curve using control points stored in a numpy array

# Create random numpy array to use as input
import numpy as np
pointPositions = np.random.uniform(-50,50,size=[15,3])

# Create curve from numpy array
curveNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsCurveNode")
slicer.util.updateMarkupsControlPointsFromArray(curveNode, pointPositions)

Add a button to module GUI to activate fiducial placement

This code snippet creates a toggle button, which activates fiducial placement when pressed (and deactivates when released).

The qSlicerMarkupsPlaceWidget widget can automatically activate placement of multiple points and can show buttons for deleting points, changing colors, lock, and hide points.

w=slicer.qSlicerMarkupsPlaceWidget()
w.setMRMLScene(slicer.mrmlScene)
markupsNodeID = slicer.modules.markups.logic().AddNewFiducialNode()
w.setCurrentNode(slicer.mrmlScene.GetNodeByID(markupsNodeID))
# Hide all buttons and only show place button
w.buttonsVisible=False
w.placeButton().show()
w.show()

Adding Fiducials via mouse clicks

You can also set the mouse mode into Markups fiducial placement by calling:

placeModePersistence = 1
slicer.modules.markups.logic().StartPlaceMode(placeModePersistence)

A lower level way to do this is via the selection and interaction nodes:

selectionNode = slicer.mrmlScene.GetNodeByID("vtkMRMLSelectionNodeSingleton")
selectionNode.SetReferenceActivePlaceNodeClassName("vtkMRMLMarkupsFiducialNode")
interactionNode = slicer.mrmlScene.GetNodeByID("vtkMRMLInteractionNodeSingleton")
placeModePersistence = 1
interactionNode.SetPlaceModePersistence(placeModePersistence)
# mode 1 is Place, can also be accessed via slicer.vtkMRMLInteractionNode().Place
interactionNode.SetCurrentInteractionMode(1)

To switch back to view transform once you’re done placing fiducials:

interactionNode = slicer.mrmlScene.GetNodeByID("vtkMRMLInteractionNodeSingleton")
interactionNode.SwitchToViewTransformMode()
# also turn off place mode persistence if required
interactionNode.SetPlaceModePersistence(0)

Access to Fiducial Properties

Each vtkMRMLMarkupsFiducialNode has a vector of points in it which can be accessed from python:

fidNode = getNode("vtkMRMLMarkupsFiducialNode1")
n = fidNode.AddFiducial(4.0, 5.5, -6.0)
fidNode.SetNthFiducialLabel(n, "new label")
# each markup is given a unique id which can be accessed from the superclass level
id1 = fidNode.GetNthMarkupID(n)
# manually set the position
fidNode.SetNthFiducialPosition(n, 6.0, 7.0, 8.0)
# set the label
fidNode.SetNthFiducialLabel(n, "New label")
# set the selected flag, only selected = 1 fiducials will be passed to CLIs
fidNode.SetNthFiducialSelected(n, 1)
# set the visibility flag
fidNode.SetNthFiducialVisibility(n, 0)

You can loop over the fiducials in a list and get the coordinates:

fidList = slicer.util.getNode("F")
numFids = fidList.GetNumberOfFiducials()
for i in range(numFids):
  ras = [0,0,0]
  fidList.GetNthFiducialPosition(i,ras)
  # the world position is the RAS position with any transform matrices applied
  world = [0,0,0,0]
  fidList.GetNthFiducialWorldCoordinates(0,world)
  print(i,": RAS =",ras,", world =",world)

You can also look at the sample code in the Endoscopy module to see how python is used to access fiducials from a scripted module.

Define/edit a circular region of interest in a slice viewer

Drop two markup points on a slice view and copy-paste the code below into the Python console. After this, as you move the markups you’ll see a circle following the markups.

# Update the sphere from the fiducial points
def UpdateSphere(param1, param2):
  """Update the sphere from the fiducial points
  """
  import math
  centerPointCoord = [0.0, 0.0, 0.0]
  markups.GetNthFiducialPosition(0,centerPointCoord)
  circumferencePointCoord = [0.0, 0.0, 0.0]
  markups.GetNthFiducialPosition(1,circumferencePointCoord)
  sphere.SetCenter(centerPointCoord)
  radius=math.sqrt((centerPointCoord[0]-circumferencePointCoord[0])**2+(centerPointCoord[1]-circumferencePointCoord[1])**2+(centerPointCoord[2]-circumferencePointCoord[2])**2)
  sphere.SetRadius(radius)
  sphere.SetPhiResolution(30)
  sphere.SetThetaResolution(30)
  sphere.Update()

# Get markup node from scene
markups=slicer.util.getNode("F")
sphere = vtk.vtkSphereSource()
UpdateSphere(0,0)

# Create model node and add to scene
modelsLogic = slicer.modules.models.logic()
model = modelsLogic.AddModel(sphere.GetOutput())
model.GetDisplayNode().SetSliceIntersectionVisibility(True)
model.GetDisplayNode().SetSliceIntersectionThickness(3)
model.GetDisplayNode().SetColor(1,1,0)

# Call UpdateSphere whenever the fiducials are changed
markups.AddObserver(slicer.vtkMRMLMarkupsNode.PointModifiedEvent, UpdateSphere, 2)

Specify a sphere by multiple of markups points

Drop multiple markup points at the boundary of the spherical object and and copy-paste the code below into the Python console to get best-fit sphere. Minimum 4 points are required, it is recommended to place the points far from each other for most accurate fit.

# Get markup node from scene
markups = slicer.util.getNode("F")

from scipy.optimize import least_squares
import numpy

def fit_sphere_least_squares(x_values, y_values, z_values, initial_parameters, bounds=((-numpy.inf, -numpy.inf, -numpy.inf, -numpy.inf),(numpy.inf, numpy.inf, numpy.inf, numpy.inf))):
    """
    Source: https://github.com/thompson318/scikit-surgery-sphere-fitting/blob/master/sksurgeryspherefitting/algorithms/sphere_fitting.py
    Uses scipy's least squares optimisor to fit a sphere to a set
    of 3D Points
    :return: x: an array containing the four fitted parameters
    :return: ier: int An integer flag. If it is equal to 1, 2, 3 or 4, the
             solution was found.
    :param: (x,y,z) three arrays of equal length containing the x, y, and z
            coordinates.
    :param: an array containing four initial values (centre, and radius)
    """
    return least_squares(_calculate_residual_sphere, initial_parameters, bounds=bounds, method="trf", jac="3-point", args=(x_values, y_values, z_values))

def _calculate_residual_sphere(parameters, x_values, y_values, z_values):
    """
    Source: https://github.com/thompson318/scikit-surgery-sphere-fitting/blob/master/sksurgeryspherefitting/algorithms/sphere_fitting.py
    Calculates the residual error for an x,y,z coordinates, fitted
    to a sphere with centre and radius defined by the parameters tuple
    :return: The residual error
    :param: A tuple of the parameters to be optimised, should contain [x_centre, y_centre, z_centre, radius]
    :param: arrays containing the x,y, and z coordinates.
    """
    #extract the parameters
    x_centre, y_centre, z_centre, radius = parameters
    #use numpy's sqrt function here, which works by element on arrays
    distance_from_centre = numpy.sqrt((x_values - x_centre)**2 + (y_values - y_centre)**2 + (z_values - z_centre)**2)
    return distance_from_centre - radius

# Fit a sphere to the markups fidicual points
markupsPositions = slicer.util.arrayFromMarkupsControlPoints(markups)
import numpy as np
# initial guess
center0 = np.mean(markupsPositions, 0)
radius0 = np.linalg.norm(np.amin(markupsPositions,0)-np.amax(markupsPositions,0))/2.0
fittingResult = fit_sphere_least_squares(markupsPositions[:,0], markupsPositions[:,1], markupsPositions[:,2], [center0[0], center0[1], center0[2], radius0])
[centerX, centerY, centerZ, radius] = fittingResult["x"]

# Create a sphere using the fitted parameters
sphere = vtk.vtkSphereSource()
sphere.SetPhiResolution(30)
sphere.SetThetaResolution(30)
sphere.SetCenter(centerX, centerY, centerZ)
sphere.SetRadius(radius)
sphere.Update()

# Add the sphere to the scene
modelsLogic = slicer.modules.models.logic()
model = modelsLogic.AddModel(sphere.GetOutput())
model.GetDisplayNode().SetSliceIntersectionVisibility(True)
model.GetDisplayNode().SetSliceIntersectionThickness(3)
model.GetDisplayNode().SetColor(1,1,0)

Measure angle between two markup planes

Measure angle between two markup plane nodes. Whenever any of the plane nodes are moved, the updated angle is printed on the console.

planeNodeNames = ["P", "P_1"]

# Print angles between slice nodes
def ShowAngle(unused1=None, unused2=None):
  planeNormalVectors = []
  for planeNodeName in planeNodeNames:
    planeNode = slicer.util.getFirstNodeByClassByName("vtkMRMLMarkupsPlaneNode", planeNodeName)
    planeNormalVector = [0.0, 0.0, 0.0]
    planeNode.GetNormalWorld(planeNormalVector)
    planeNormalVectors.append(planeNormalVector)
  angleRad = vtk.vtkMath.AngleBetweenVectors(planeNormalVectors[0], planeNormalVectors[1])
  angleDeg = vtk.vtkMath.DegreesFromRadians(angleRad)
  print("Angle between planes {0} and {1} = {2:0.3f}".format(planeNodeNames[0], planeNodeNames[1], angleDeg))

# Observe plane node changes
for planeNodeName in planeNodeNames:
  planeNode = slicer.util.getFirstNodeByClassByName("vtkMRMLMarkupsPlaneNode", planeNodeName)
  planeNode.AddObserver(slicer.vtkMRMLMarkupsPlaneNode.PointModifiedEvent, ShowAngle)

# Print current angle
ShowAngle()

Measure angle between two markup lines

Measure angle between two markup line nodes. Whenever either line is moved, the updated angle is printed on the console.

lineNodeNames = ["L", "L_1"]

# Print angles between slice nodes
def ShowAngle(unused1=None, unused2=None):
  import numpy as np
  lineDirectionVectors = []
  for lineNodeName in lineNodeNames:
    lineNode = slicer.util.getFirstNodeByClassByName("vtkMRMLMarkupsLineNode", lineNodeName)
    lineStartPos = np.zeros(3)
    lineEndPos = np.zeros(3)
    lineNode.GetNthControlPointPositionWorld(0, lineStartPos)
    lineNode.GetNthControlPointPositionWorld(1, lineEndPos)
    lineDirectionVector = (lineEndPos-lineStartPos)/np.linalg.norm(lineEndPos-lineStartPos)
    lineDirectionVectors.append(lineDirectionVector)
  angleRad = vtk.vtkMath.AngleBetweenVectors(lineDirectionVectors[0], lineDirectionVectors[1])
  angleDeg = vtk.vtkMath.DegreesFromRadians(angleRad)
  print("Angle between lines {0} and {1} = {2:0.3f}".format(lineNodeNames[0], lineNodeNames[1], angleDeg))

# Observe line node changes
for lineNodeName in lineNodeNames:
  lineNode = slicer.util.getFirstNodeByClassByName("vtkMRMLMarkupsLineNode", lineNodeName)
  lineNode.AddObserver(slicer.vtkMRMLMarkupsLineNode.PointModifiedEvent, ShowAngle)

# Print current angle
ShowAngle()

Set slice position and orientation from 3 markup fiducials

Drop 3 markup points in the scene and copy-paste the code below into the Python console. After this, as you move the markups you’ll see the red slice view position and orientation will be set to make it fit to the 3 points.

# Update plane from fiducial points
def UpdateSlicePlane(param1=None, param2=None):
  # Get point positions as numpy array
  import numpy as np
  nOfFiduciallPoints = markups.GetNumberOfFiducials()
  if nOfFiduciallPoints < 3:
    return  # not enough points
  points = np.zeros([3,nOfFiduciallPoints])
  for i in range(0, nOfFiduciallPoints):
    markups.GetNthFiducialPosition(i, points[:,i])
  # Compute plane position and normal
  planePosition = points.mean(axis=1)
  planeNormal = np.cross(points[:,1] - points[:,0], points[:,2] - points[:,0])
  planeX = points[:,1] - points[:,0]
  sliceNode.SetSliceToRASByNTP(planeNormal[0], planeNormal[1], planeNormal[2],
    planeX[0], planeX[1], planeX[2],
    planePosition[0], planePosition[1], planePosition[2], 0)

# Get markup node from scene
sliceNode = slicer.app.layoutManager().sliceWidget("Red").mrmlSliceNode()
markups = slicer.util.getNode("F")

# Update slice plane manually
UpdateSlicePlane()

# Update slice plane automatically whenever points are changed
markupObservation = [markups, markups.AddObserver(slicer.vtkMRMLMarkupsNode.PointModifiedEvent, UpdateSlicePlane, 2)]

To stop automatic updates, run this:

markupObservation[0].RemoveObserver(markupObservation[1])

Switching to markup fiducial placement mode

To activate a fiducial placement mode, both interaction mode has to be set and a fiducial node has to be selected:

interactionNode = slicer.app.applicationLogic().GetInteractionNode()
selectionNode = slicer.app.applicationLogic().GetSelectionNode()
selectionNode.SetReferenceActivePlaceNodeClassName("vtkMRMLMarkupsFiducialNode")
fiducialNode = slicer.vtkMRMLMarkupsFiducialNode()
slicer.mrmlScene.AddNode(fiducialNode)
fiducialNode.CreateDefaultDisplayNodes()
selectionNode.SetActivePlaceNodeID(fiducialNode.GetID())
interactionNode.SetCurrentInteractionMode(interactionNode.Place)

Alternatively, qSlicerMarkupsPlaceWidget widget can be used to initiate markup placement:

# Temporary markups node
markupsNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsFiducialNode")

def placementModeChanged(active):
  print("Placement: " +("active" if active else "inactive"))
  # You can inspect what is in the markups node here, delete the temporary markup node, etc.

# Create and set up widget that contains a single "place markup" button. The widget can be placed in the module GUI.
placeWidget = slicer.qSlicerMarkupsPlaceWidget()
placeWidget.setMRMLScene(slicer.mrmlScene)
placeWidget.setCurrentNode(markupsNode)
placeWidget.buttonsVisible=False
placeWidget.placeButton().show()
placeWidget.connect("activeMarkupsFiducialPlaceModeChanged(bool)", placementModeChanged)
placeWidget.show()

Change markup fiducial display properties

Display properties are stored in display node(s) associated with the fiducial node.

fiducialNode = getNode("F")
fiducialDisplayNode = fiducialNode.GetDisplayNode()
fiducialDisplayNode.SetVisibility(False) # Hide all points
fiducialDisplayNode.SetVisibility(True) # Show all points
fiducialDisplayNode.SetSelectedColor(1,1,0) # Set color to yellow
fiducialDisplayNode.SetViewNodeIDs(["vtkMRMLSliceNodeRed", "vtkMRMLViewNode1"]) # Only show in red slice view and first 3D view

Get a notification if a markup point position is modified

Event management of Slicer-4.11 version is still subject to change. The example below shows how point manipulation can be observed now.

def onMarkupChanged(caller,event):
  markupsNode = caller
  sliceView = markupsNode.GetAttribute("Markups.MovingInSliceView")
  movingMarkupIndex = markupsNode.GetDisplayNode().GetActiveControlPoint()
  if movingMarkupIndex >= 0:
    pos = [0,0,0]
    markupsNode.GetNthFiducialPosition(movingMarkupIndex, pos)
    isPreview = markupsNode.GetNthControlPointPositionStatus(movingMarkupIndex) == slicer.vtkMRMLMarkupsNode.PositionPreview
    if isPreview:
      logging.info("Point {0} is previewed at {1} in slice view {2}".format(movingMarkupIndex, pos, sliceView))
    else:
      logging.info("Point {0} was moved {1} in slice view {2}".format(movingMarkupIndex, pos, sliceView))
  else:
    logging.info("Points modified: slice view = {0}".format(sliceView))

def onMarkupStartInteraction(caller, event):
  markupsNode = caller
  sliceView = markupsNode.GetAttribute("Markups.MovingInSliceView")
  movingMarkupIndex = markupsNode.GetDisplayNode().GetActiveControlPoint()
  logging.info("Start interaction: point ID = {0}, slice view = {1}".format(movingMarkupIndex, sliceView))

def onMarkupEndInteraction(caller, event):
  markupsNode = caller
  sliceView = markupsNode.GetAttribute("Markups.MovingInSliceView")
  movingMarkupIndex = markupsNode.GetDisplayNode().GetActiveControlPoint()
  logging.info("End interaction: point ID = {0}, slice view = {1}".format(movingMarkupIndex, sliceView))

markupsNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsFiducialNode")
markupsNode.CreateDefaultDisplayNodes()
markupsNode.AddFiducial(0,0,0)
markupsNode.AddObserver(slicer.vtkMRMLMarkupsNode.PointModifiedEvent, onMarkupChanged)
markupsNode.AddObserver(slicer.vtkMRMLMarkupsNode.PointStartInteractionEvent, onMarkupStartInteraction)
markupsNode.AddObserver(slicer.vtkMRMLMarkupsNode.PointEndInteractionEvent, onMarkupEndInteraction)

Write markup positions to JSON file

markupNode = getNode("F")
outputFileName = "c:/tmp/test.json"

# Get markup positions
data = []
for fidIndex in range(markupNode.GetNumberOfFiducials()):
  coords=[0,0,0]
  markupNode.GetNthFiducialPosition(fidIndex,coords)
  data.append({"label": markupNode.GetNthFiducialLabel(), "position": coords})

import json
with open(outputFileName, "w") as outfile:
  json.dump(data, outfile)

Write annotation ROI to JSON file

roiNode = getNode("R")
outputFileName = "c:/tmp/test.json"

# Get annotation ROI data
center = [0,0,0]
radius = [0,0,0]
roiNode.GetControlPointWorldCoordinates(0, center)
roiNode.GetControlPointWorldCoordinates(1, radius)
data = {"center": radius, "radius": radius}

# Write to json file
import json
with open(outputFileName, "w") as outfile:
  json.dump(data, outfile)

Fit slice plane to markup fiducials

sliceNode = slicer.mrmlScene.GetNodeByID("vtkMRMLSliceNodeRed")
markupsNode = slicer.mrmlScene.GetFirstNodeByName("F")
# Get markup point positions as numpy arrays
import numpy as np
p1 = np.zeros(3)
p2 = np.zeros(3)
p3 = np.zeros(3)
markupsNode.GetNthFiducialPosition(0, p1)
markupsNode.GetNthFiducialPosition(1, p2)
markupsNode.GetNthFiducialPosition(2, p3)
# Get plane axis directions
n = np.cross(p2-p1, p2-p3) # plane normal direction
n = n/np.linalg.norm(n)
t = np.cross([0.0, 0.0, 1], n) # plane transverse direction
t = t/np.linalg.norm(t)
# Set slice plane orientation and position
sliceNode.SetSliceToRASByNTP(n[0], n[1], n[2], t[0], t[1], t[2], p1[0], p1[1], p1[2], 0)

Change color of a markups node

Markups have Color and SelectedColor properties. SelectedColor is used if all control points are in “selected” state, which is the default. So, in most cases SetSelectedColor method must be used to change markups node color.

Display list of control points in my module’s GUI

The qSlicerSimpleMarkupsWidget can be integrated into module widgets to display list of markups control points and initiate placement. An example of this use is in Gel Dosimetry module.

Pre-populate the scene with measurements

This code snippet creates a set of predefined line markups (named A, B, C, D) in the scene when the user hits Ctrl+N. How to use this:

  1. Customize the code (replace A, B, C, D with your measurement names) and copy-paste the code into the Python console. This has to be done only once after Slicer is started. Add it to .slicerrc.py file so that it persists even if Slicer is restarted.
  2. Load the data set that has to be measured
  3. Hit Ctrl+N to create all the measurements
  4. Go to Markups module to see the list of measurements
  5. For each measurement: select it in the data tree, click on the place button on the toolbar then click in slice or 3D views
sliceNode = slicer.mrmlScene.GetNodeByID("vtkMRMLSliceNodeRed")
def createMeasurements():
  for nodeName in ['A', 'B', 'C', 'D']:
    lineNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsLineNode", nodeName)
    lineNode.CreateDefaultDisplayNodes()
    dn = lineNode.GetDisplayNode()
    # Use crosshair glyph to allow more accurate point placement
    dn.SetGlyphTypeFromString("CrossDot2D")
    # Hide measurement result while markup up
    lineNode.GetMeasurement('length').SetEnabled(False)

shortcut1 = qt.QShortcut(slicer.util.mainWindow())
shortcut1.setKey(qt.QKeySequence("Ctrl+n"))
shortcut1.connect( 'activated()', createMeasurements)

Copy all measurements in the scene to Excel

This code snippet creates a set of predefined line markups (named A, B, C, D) in the scene when the user hits Ctrl+N. How to use this:

  1. Copy-paste the code into the Python console. This has to be done only once after Slicer is started. Add it to .slicerrc.py file so that it persists even if Slicer is restarted.
  2. Load the data set that has to be measured and place line markups (you can use the “Pre-populate the scene with measurements” script above to help with this)
  3. Hit Ctrl+M to copy all line measurents to the clipboard
  4. Switch to Excel and hit Ctrl+V to paste the results there
  5. Save the scene, just in case later you need to review your measurements
def copyLineMeasurementsToClipboard():
  measurements = []
  # Collect all line measurements from the scene
  lineNodes = getNodesByClass('vtkMRMLMarkupsLineNode')
  for lineNode in lineNodes:
    # Get node filename that the length was measured on
    try:
      volumeNode = slicer.mrmlScene.GetNodeByID(lineNode.GetNthMarkupAssociatedNodeID(0))
      imagePath = volumeNode.GetStorageNode().GetFileName()
    except:
      imagePath = ''
    # Get line node n
    measurementName = lineNode.GetName()
    # Get length measurement
    lineNode.GetMeasurement('length').SetEnabled(True)
    length = str(lineNode.GetMeasurement('length').GetValue())
    # Add fields to results
    measurements.append('\t'.join([imagePath, measurementName, length]))
  # Copy all measurements to clipboard (to be pasted into Excel)
  slicer.app.clipboard().setText("\n".join(measurements))
  slicer.util.delayDisplay(f"Copied {len(measurements)} length measurements to the clipboard.")

shortcut2 = qt.QShortcut(slicer.util.mainWindow())
shortcut2.setKey(qt.QKeySequence("Ctrl+m"))
shortcut2.connect( 'activated()', copyLineMeasurementsToClipboard)

Use markups json files in Python - outside Slicer

The examples below show how to use markups json files outside Slicer, in any Python environment.

To access content of a json file it can be either read as a json document or directly into a pandas dataframe using a single command.

Get a table of control point labels and positions

Get table from the first markups node in the file:

import pandas as pd
controlPointsTable = pd.DataFrame.from_dict(pd.read_json(input_json_filename)['markups'][0]['controlPoints'])

Result:

>>> controlPointsTable
  label                                        position
0   F-1  [-53.388409961685824, -73.33572796934868, 0.0]
1   F-2     [49.8682950191571, -88.58955938697324, 0.0]
2   F-3   [-25.22749042145594, 59.255268199233726, 0.0]

Access position of control points positions in separate x, y, z columns

controlPointsTable[['x','y','z']] = pd.DataFrame(controlPointsTable['position'].to_list())
del controlPointsTable['position']

Write control points to a csv file

controlPointsTable.to_csv(output_csv_filename)

Resulting csv file:

,label,x,y,z
0,F-1,-53.388409961685824,-73.33572796934868,0.0
1,F-2,49.8682950191571,-88.58955938697324,0.0
2,F-3,-25.22749042145594,59.255268199233726,0.0

Assign custom actions to markups

Custom actions can be assigned to markups, which can be triggered by any interaction event (mouse or keyboard action). The actions can be detected by adding observers to the markup node’s display node.

# This example adds an action to the default double-click action on a markup
# and defines two new custom actions. It is done for all existing markups in the first 3D view.
#
# How to use:
# 1. Create markups nodes.
# 2. Run the script below.
# 3. Double-click on the markup -> this triggers toggleLabelVisibilty.
# 4. Hover the mouse over a markup then pressing `q` and `w` keys -> this triggers shrinkControlPoints and growControlPoints.

threeDViewWidget = slicer.app.layoutManager().threeDWidget(0)
markupsDisplayableManager = threeDViewWidget.threeDView().displayableManagerByClassName('vtkMRMLMarkupsDisplayableManager')

def shrinkControlPoints(caller, eventId):
    markupsDisplayNode = caller
    markupsDisplayNode.SetGlyphScale(markupsDisplayNode.GetGlyphScale()/1.1)

def growControlPoints(caller, eventId):
    markupsDisplayNode = caller
    markupsDisplayNode.SetGlyphScale(markupsDisplayNode.GetGlyphScale()*1.1)

def toggleLabelVisibility(caller, eventId):
    markupsDisplayNode = caller
    markupsDisplayNode.SetPointLabelsVisibility(not markupsDisplayNode.GetPointLabelsVisibility())

observations = []  # store the observations so that later can be removed
markupsDisplayNodes = slicer.util.getNodesByClass("vtkMRMLMarkupsDisplayNode")
for markupsDisplayNode in markupsDisplayNodes:
    # Assign keyboard shortcut to trigger custom actions
    markupsWidget = markupsDisplayableManager.GetWidget(markupsDisplayNode)
    # Left double-click interaction event is translated to markupsWidget.WidgetEventAction by default,
    # therefore we don't need to add an event translation for that. We just add two keyboard event translation for two custom actions
    markupsWidget.SetKeyboardEventTranslation(markupsWidget.WidgetStateOnWidget, vtk.vtkEvent.NoModifier, '\0', 0, "q", markupsWidget.WidgetEventCustomAction1)
    markupsWidget.SetKeyboardEventTranslation(markupsWidget.WidgetStateOnWidget, vtk.vtkEvent.NoModifier, '\0', 0, "w", markupsWidget.WidgetEventCustomAction2)
    # Add observer to custom actions
    observations.append([markupsDisplayNode, markupsDisplayNode.AddObserver(markupsDisplayNode.ActionEvent, toggleLabelVisibility)])
    observations.append([markupsDisplayNode, markupsDisplayNode.AddObserver(markupsDisplayNode.CustomActionEvent1, shrinkControlPoints)])
    observations.append([markupsDisplayNode, markupsDisplayNode.AddObserver(markupsDisplayNode.CustomActionEvent2, growControlPoints)])

Remove observations when custom actions are not needed anymore by uncommenting these lines:

for observedNode, observation in observations:
    observedNode.RemoveObserver(observation)

Models

Show a simple surface mesh as a model node

This example shows how to display a simple surface mesh (a box, created by a VTK source filter) as a model node.

# Create and set up polydata source
box = vtk.vtkCubeSource()
box.SetXLength(30)
box.SetYLength(20)
box.SetZLength(15)
box.SetCenter(10,20,5)

# Create a model node that displays output of the source
boxNode = slicer.modules.models.logic().AddModel(box.GetOutputPort())

# Adjust display properties
boxNode.GetDisplayNode().SetColor(1,0,0)
boxNode.GetDisplayNode().SetOpacity(0.8)

Measure distance of points from surface

This example computes closest distance of points (markups fiducial F) from a surface (model node mymodel) and writes results into a table.

markupsNode = getNode("F")
modelNode = getNode("mymodel")

# Transform model polydata to world coordinate system
if modelNode.GetParentTransformNode():
  transformModelToWorld = vtk.vtkGeneralTransform()
  slicer.vtkMRMLTransformNode.GetTransformBetweenNodes(modelNode.GetParentTransformNode(), None, transformModelToWorld)
  polyTransformToWorld = vtk.vtkTransformPolyDataFilter()
  polyTransformToWorld.SetTransform(transformModelToWorld)
  polyTransformToWorld.SetInputData(modelNode.GetPolyData())
  polyTransformToWorld.Update()
  surface_World = polyTransformToWorld.GetOutput()
else:
  surface_World = modelNode.GetPolyData()

# Create arrays to store results
indexCol = vtk.vtkIntArray()
indexCol.SetName("Index")
labelCol = vtk.vtkStringArray()
labelCol.SetName("Name")
distanceCol = vtk.vtkDoubleArray()
distanceCol.SetName("Distance")

distanceFilter = vtk.vtkImplicitPolyDataDistance()
distanceFilter.SetInput(surface_World);
nOfFiduciallPoints = markupsNode.GetNumberOfFiducials()
for i in range(0, nOfFiduciallPoints):
  point_World = [0,0,0]
  markupsNode.GetNthControlPointPositionWorld(i, point_World)
  closestPointOnSurface_World = [0,0,0]
  closestPointDistance = distanceFilter.EvaluateFunctionAndGetClosestPoint(point_World, closestPointOnSurface_World)
  indexCol.InsertNextValue(i)
  labelCol.InsertNextValue(markupsNode.GetNthControlPointLabel(i))
  distanceCol.InsertNextValue(closestPointDistance)

# Create a table from result arrays
resultTableNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLTableNode", "Points from surface distance")
resultTableNode.AddColumn(indexCol)
resultTableNode.AddColumn(labelCol)
resultTableNode.AddColumn(distanceCol)

# Show table in view layout
slicer.app.layoutManager().setLayout(slicer.vtkMRMLLayoutNode.SlicerLayoutFourUpTableView)
slicer.app.applicationLogic().GetSelectionNode().SetReferenceActiveTableID(resultTableNode.GetID())
slicer.app.applicationLogic().PropagateTableSelection()

Add a texture mapped plane to the scene as a model

Note that model textures are not exposed in the GUI and are not saved in the scene

# Create model node
planeSource = vtk.vtkPlaneSource()
planeSource.SetOrigin(-50.0, -50.0, 0.0)
planeSource.SetPoint1(50.0, -50.0, 0.0)
planeSource.SetPoint2(-50.0, 50.0, 0.0)
model = slicer.modules.models.logic().AddModel(planeSource.GetOutputPort())

# Tune display properties
modelDisplay = model.GetDisplayNode()
modelDisplay.SetColor(1,1,0) # yellow
modelDisplay.SetBackfaceCulling(0)

# Add texture (just use image of an ellipsoid)
e = vtk.vtkImageEllipsoidSource()
modelDisplay.SetTextureImageDataConnection(e.GetOutputPort())

Get scalar values at surface of a model

The following script allows getting selected scalar value at a selected position of a model. Position can be selected by moving the mouse while holding down Shift key.

modelNode = getNode("sphere")
modelPointValues = modelNode.GetPolyData().GetPointData().GetArray("Normals")
markupsNode = slicer.mrmlScene.GetFirstNodeByName("F")

if not markupsNode:
  markupsNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsFiducialNode","F")

pointsLocator = vtk.vtkPointLocator() # could try using vtk.vtkStaticPointLocator() if need to optimize
pointsLocator.SetDataSet(modelNode.GetPolyData())
pointsLocator.BuildLocator()

def onMouseMoved(observer,eventid):
  ras=[0,0,0]
  crosshairNode.GetCursorPositionRAS(ras)
  if markupsNode.GetNumberOfFiducials() == 0:
    markupsNode.AddFiducial(*ras)
  else:
    markupsNode.SetNthFiducialPosition(0,*ras)
  closestPointId = pointsLocator.FindClosestPoint(ras)
  closestPointValue = modelPointValues.GetTuple(closestPointId)
  print("RAS = " + repr(ras) + "    value = " + repr(closestPointValue))

crosshairNode=slicer.util.getNode("Crosshair")
observationId = crosshairNode.AddObserver(slicer.vtkMRMLCrosshairNode.CursorPositionModifiedEvent, onMouseMoved)

# To stop printing of values run this:
# crosshairNode.RemoveObserver(observationId)

Apply VTK filter on a model node

modelNode = getNode("tip")

# Compute curvature
curv = vtk.vtkCurvatures()
curv.SetInputData(modelNode.GetPolyData())
modelNode.SetPolyDataConnection(curv.GetOutputPort())

# Set up coloring by Curvature
modelNode.GetDisplayNode().SetActiveScalar("Gauss_Curvature", vtk.vtkAssignAttribute.POINT_DATA)
modelNode.GetDisplayNode().SetAndObserveColorNodeID("Viridis")
modelNode.GetDisplayNode().SetScalarVisibility(True)

Select cells of a model using markups fiducial points

The following script selects cells of a model node that are closest to positions of markups fiducial points.

# Get input nodes
modelNode = slicer.util.getNode("Segment_1") # select cells in this model
markupsNode = slicer.util.getNode("F") # points will be selected at positions specified by this markups fiducial node

# Create scalar array that will store selection state
cellScalars = modelNode.GetMesh().GetCellData()
selectionArray = cellScalars.GetArray("selection")
if not selectionArray:
  selectionArray = vtk.vtkIntArray()
  selectionArray.SetName("selection")
  selectionArray.SetNumberOfValues(modelNode.GetMesh().GetNumberOfCells())
  selectionArray.Fill(0)
  cellScalars.AddArray(selectionArray)

# Set up coloring by selection array
modelNode.GetDisplayNode().SetActiveScalar("selection", vtk.vtkAssignAttribute.CELL_DATA)
modelNode.GetDisplayNode().SetAndObserveColorNodeID("vtkMRMLColorTableNodeWarm1")
modelNode.GetDisplayNode().SetScalarVisibility(True)

# Initialize cell locator
cell = vtk.vtkCellLocator()
cell.SetDataSet(modelNode.GetMesh())
cell.BuildLocator()

def onPointsModified(observer=None, eventid=None):
  global markupsNode, selectionArray
  selectionArray.Fill(0) # set all cells to non-selected by default
  markupPoints = slicer.util.arrayFromMarkupsControlPoints(markupsNode)
  closestPoint = [0.0, 0.0, 0.0]
  cellObj = vtk.vtkGenericCell()
  cellId = vtk.mutable(0)
  subId = vtk.mutable(0)
  dist2 = vtk.mutable(0.0)
  for markupPoint in markupPoints:
    cell.FindClosestPoint(markupPoint, closestPoint, cellObj, cellId, subId, dist2)
    closestCell = cellId.get()
    if closestCell >=0:
      selectionArray.SetValue(closestCell, 100) # set selected cell's scalar value to non-zero
  selectionArray.Modified()

# Initial update
onPointsModified()
# Automatic update each time when a markup point is modified
markupsNodeObserverTag = markupsNode.AddObserver(slicer.vtkMRMLMarkupsFiducialNode.PointModifiedEvent, onPointsModified)

# To stop updating selection, run this:
# markupsNode.RemoveObserver(markupsNodeObserverTag)

Export entire scene as VRML

Save all surface meshes displayed in the scene (models, markups, etc). Solid colors and coloring by scalar is preserved. Textures are not supported.

exporter = vtk.vtkVRMLExporter()
exporter.SetRenderWindow(slicer.app.layoutManager().threeDWidget(0).threeDView().renderWindow())
exporter.SetFileName("C:/tmp/something.wrl")
exporter.Write()

Export model to Blender, including color by scalar

modelNode = getNode("Model")
plyFilePath = "c:/tmp/model.ply"

modelDisplayNode = modelNode.GetDisplayNode()
triangles = vtk.vtkTriangleFilter()
triangles.SetInputConnection(modelDisplayNode.GetOutputPolyDataConnection())

plyWriter = vtk.vtkPLYWriter()
plyWriter.SetInputConnection(triangles.GetOutputPort())
lut = vtk.vtkLookupTable()
lut.DeepCopy(modelDisplayNode.GetColorNode().GetLookupTable())
lut.SetRange(modelDisplayNode.GetScalarRange())
plyWriter.SetLookupTable(lut)
plyWriter.SetArrayName(modelDisplayNode.GetActiveScalarName())

plyWriter.SetFileName(plyFilePath)
plyWriter.Write()

Show comparison view of all model files a folder

# Inputs
modelDir = "c:/some/folder/containing/models"
modelFileExt = "stl"
numberOfColumns = 4

import math
import os
modelFiles = list(f for f in os.listdir(modelDir) if f.endswith("." + modelFileExt))

# Create a custom layout
numberOfRows = int(math.ceil(len(modelFiles)/numberOfColumns))
customLayoutId=567  # we pick a random id that is not used by others
slicer.app.setRenderPaused(True)
customLayout = '<layout type="vertical">'
viewIndex = 0
for rowIndex in range(numberOfRows):
  customLayout += '<item><layout type="horizontal">'
  for colIndex in range(numberOfColumns):
    name = os.path.basename(modelFiles[viewIndex]) if viewIndex < len(modelFiles) else "compare " + str(viewIndex)
    customLayout += '<item><view class="vtkMRMLViewNode" singletontag="'+name
    customLayout += '"><property name="viewlabel" action="default">'+name+'</property></view></item>'
    viewIndex += 1
  customLayout += '</layout></item>'

customLayout += '</layout>'
if not slicer.app.layoutManager().layoutLogic().GetLayoutNode().SetLayoutDescription(customLayoutId, customLayout):
    slicer.app.layoutManager().layoutLogic().GetLayoutNode().AddLayoutDescription(customLayoutId, customLayout)

slicer.app.layoutManager().setLayout(customLayoutId)

# Load and show each model in a view
for modelIndex, modelFile in enumerate(modelFiles):
  # Show only one model in each view
  name = os.path.basename(modelFile)
  viewNode = slicer.mrmlScene.GetSingletonNode(name, "vtkMRMLViewNode")
  viewNode.LinkedControlOn()
  modelNode = slicer.util.loadModel(modelDir + "/" + modelFile)
  modelNode.GetDisplayNode().AddViewNodeID(viewNode.GetID())

slicer.app.setRenderPaused(False)

Rasterize a model and save it to a series of image files

This example shows how to generate a stack of image files from an STL file:

inputModelFile = "/some/input/folder/SomeShape.stl"
outputDir = "/some/output/folder"
outputVolumeLabelValue = 100
outputVolumeSpacingMm = [0.5, 0.5, 0.5]
outputVolumeMarginMm = [10.0, 10.0, 10.0]
# Read model
inputModel = slicer.util.loadModel(inputModelFile)
# Determine output volume geometry and create a corresponding reference volume
import math
import numpy as np
bounds = np.zeros(6)
inputModel.GetBounds(bounds)
imageData = vtk.vtkImageData()
imageSize = [ int((bounds[axis*2+1]-bounds[axis*2]+outputVolumeMarginMm[axis]*2.0)/outputVolumeSpacingMm[axis]) for axis in range(3) ]
imageOrigin = [ bounds[axis*2]-outputVolumeMarginMm[axis] for axis in range(3) ]
imageData.SetDimensions(imageSize)
imageData.AllocateScalars(vtk.VTK_UNSIGNED_CHAR, 1)
imageData.GetPointData().GetScalars().Fill(0)
referenceVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode")
referenceVolumeNode.SetOrigin(imageOrigin)
referenceVolumeNode.SetSpacing(outputVolumeSpacingMm)
referenceVolumeNode.SetAndObserveImageData(imageData)
referenceVolumeNode.CreateDefaultDisplayNodes()
# Convert model to labelmap
seg = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLSegmentationNode")
seg.SetReferenceImageGeometryParameterFromVolumeNode(referenceVolumeNode)
slicer.modules.segmentations.logic().ImportModelToSegmentationNode(inputModel, seg)
seg.CreateBinaryLabelmapRepresentation()
outputLabelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportVisibleSegmentsToLabelmapNode(seg, outputLabelmapVolumeNode, referenceVolumeNode)
outputLabelmapVolumeArray = (slicer.util.arrayFromVolume(outputLabelmapVolumeNode) * outputVolumeLabelValue).astype("int8")
# Write labelmap volume to series of TIFF files
pip_install("imageio")
import imageio
for i in range(len(outputLabelmapVolumeArray)):
  imageio.imwrite(f"{outputDir}/image_{i:03}.tiff", outputLabelmapVolumeArray[i])

Plots

Slicer plots displayed in view layout

Create histogram plot of a volume and show it embedded in the view layout. More information: https://www.slicer.org/wiki/Documentation/Nightly/Developers/Plots

Using slicer.util.plot() utility function

# Get a volume from SampleData and compute its histogram
import SampleData
import numpy as np
volumeNode = SampleData.SampleDataLogic().downloadMRHead()
histogram = np.histogram(arrayFromVolume(volumeNode), bins=50)

chartNode = slicer.util.plot(histogram, xColumnIndex = 1)
chartNode.SetYAxisRangeAuto(False)
chartNode.SetYAxisRange(0, 4e5)
https://www.slicer.org/w/img_auth.php/9/9c/SlicerPlot.png

Plot displayed using Slicer’s plotting module

Using MRML classes only

# Get a volume from SampleData
import SampleData
volumeNode = SampleData.SampleDataLogic().downloadMRHead()

# Compute histogram values
import numpy as np
histogram = np.histogram(arrayFromVolume(volumeNode), bins=50)

# Save results to a new table node
tableNode=slicer.mrmlScene.AddNewNodeByClass("vtkMRMLTableNode")
updateTableFromArray(tableNode, histogram)
tableNode.GetTable().GetColumn(0).SetName("Count")
tableNode.GetTable().GetColumn(1).SetName("Intensity")

# Create plot
plotSeriesNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLPlotSeriesNode", volumeNode.GetName() + " histogram")
plotSeriesNode.SetAndObserveTableNodeID(tableNode.GetID())
plotSeriesNode.SetXColumnName("Intensity")
plotSeriesNode.SetYColumnName("Count")
plotSeriesNode.SetPlotType(plotSeriesNode.PlotTypeScatterBar)
plotSeriesNode.SetColor(0, 0.6, 1.0)

# Create chart and add plot
plotChartNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLPlotChartNode")
plotChartNode.AddAndObservePlotSeriesNodeID(plotSeriesNode.GetID())
plotChartNode.YAxisRangeAutoOff()
plotChartNode.SetYAxisRange(0, 500000)

# Show plot in layout
slicer.modules.plots.logic().ShowChartInLayout(plotChartNode)

Using matplotlib

Matplotlib may be used from within Slicer, but the default Tk backend locks up and crashes Slicer. However, Matplotlib may still be used through other backends. More details can be found on the MatPlotLib pages.

Non-interactive plot

try:
  import matplotlib
except ModuleNotFoundError:
  pip_install("matplotlib")
  import matplotlib

matplotlib.use("Agg")
from pylab import *

t1 = arange(0.0, 5.0, 0.1)
t2 = arange(0.0, 5.0, 0.02)
t3 = arange(0.0, 2.0, 0.01)

subplot(211)
plot(t1, cos(2*pi*t1)*exp(-t1), "bo", t2, cos(2*pi*t2)*exp(-t2), "k")
grid(True)
title("A tale of 2 subplots")
ylabel("Damped")

subplot(212)
plot(t3, cos(2*pi*t3), "r--")
grid(True)
xlabel("time (s)")
ylabel("Undamped")
savefig("MatplotlibExample.png")

# Static image view
pm = qt.QPixmap("MatplotlibExample.png")
imageWidget = qt.QLabel()
imageWidget.setPixmap(pm)
imageWidget.setScaledContents(True)
imageWidget.show()
https://www.slicer.org/w/img_auth.php/a/ab/MatplotlibExample.png

Matplotlib example

Plot in Slicer Jupyter notebook

import JupyterNotebooksLib as slicernb
try:
  import matplotlib
except ModuleNotFoundError:
  pip_install("matplotlib")
  import matplotlib

matplotlib.use("Agg")

import matplotlib.pyplot as plt
import numpy as np

def f(t):
  s1 = np.cos(2*np.pi*t)
  e1 = np.exp(-t)
  return s1 * e1

t1 = np.arange(0.0, 5.0, 0.1)
t2 = np.arange(0.0, 5.0, 0.02)
t3 = np.arange(0.0, 2.0, 0.01)


fig, axs = plt.subplots(2, 1, constrained_layout=True)
axs[0].plot(t1, f(t1), "o", t2, f(t2), "-")
axs[0].set_title("subplot 1")
axs[0].set_xlabel("distance (m)")
axs[0].set_ylabel("Damped oscillation")
fig.suptitle("This is a somewhat long figure title", fontsize=16)

axs[1].plot(t3, np.cos(2*np.pi*t3), "--")
axs[1].set_xlabel("time (s)")
axs[1].set_title("subplot 2")
axs[1].set_ylabel("Undamped")

slicernb.MatplotlibDisplay(matplotlib.pyplot)
https://www.slicer.org/w/img_auth.php/a/a2/JupyterNotebookMatplotlibExample.png

Example for using Matplotlib in a Slicer Jupyter Notebook

Interactive plot using wxWidgets GUI toolkit

try:
  import matplotlib
  import wx
except ModuleNotFoundError:
  pip_install("matplotlib wxPython")
  import matplotlib

# Get a volume from SampleData and compute its histogram
import SampleData
import numpy as np
volumeNode = SampleData.SampleDataLogic().downloadMRHead()
histogram = np.histogram(arrayFromVolume(volumeNode), bins=50)

# Set matplotlib to use WXAgg backend
import matplotlib
matplotlib.use("WXAgg")

# Show an interactive plot
import matplotlib.pyplot as plt
fig, ax = plt.subplots()
ax.plot(histogram[1][1:], histogram[0].astype(float))
ax.grid(True)
ax.set_ylim((0, 4e5))
plt.show(block=False)
https://www.slicer.org/w/img_auth.php/d/d2/InteractiveMatplotlibExample.png

Interactive Matplotlib Example

Screen Capture

Capture the full Slicer screen and save it into a file

img = qt.QPixmap.grabWidget(slicer.util.mainWindow()).toImage()
img.save("c:/tmp/test.png")

Capture all the views save it into a file

import ScreenCapture
cap = ScreenCapture.ScreenCaptureLogic()
cap.showViewControllers(False)
cap.captureImageFromView(None, "c:/tmp/test.png")
cap.showViewControllers(True)

Capture a single view

viewNodeID = "vtkMRMLViewNode1"
import ScreenCapture
cap = ScreenCapture.ScreenCaptureLogic()
view = cap.viewFromNode(slicer.mrmlScene.GetNodeByID(viewNodeID))
cap.captureImageFromView(view, "c:/tmp/test.png")

Common values for viewNodeID: vtkMRMLSliceNodeRed, vtkMRMLSliceNodeYellow, vtkMRMLSliceNodeGreen, vtkMRMLViewNode1, vtkMRMLViewNode2. The ScreenCapture module can also create video animations of rotating views, slice sweeps, etc.

Capture a slice view sweep into a series of PNG files

For example, Red slice view, 30 images, from position -125.0 to 75.0, into c:/tmp folder, with name image_00001.png, image_00002.png, …

import ScreenCapture
ScreenCapture.ScreenCaptureLogic().captureSliceSweep(getNode("vtkMRMLSliceNodeRed"), -125.0, 75.0, 30, "c:/tmp", "image_%05d.png")

Capture 3D view into PNG file with transparent background

renderWindow = slicer.app.layoutManager().threeDWidget(0).threeDView().renderWindow()
renderWindow.SetAlphaBitPlanes(1)
wti = vtk.vtkWindowToImageFilter()
wti.SetInputBufferTypeToRGBA()
wti.SetInput(renderWindow)
writer = vtk.vtkPNGWriter()
writer.SetFileName("c:/tmp/screenshot.png")
writer.SetInputConnection(wti.GetOutputPort())
writer.Write()

Save a series of images from a slice view

You can use ScreenCapture module to capture series of images. To do it programmatically, save the following into a file such as /tmp/record.py and then in the slicer python console type execfile("/tmp/record.py")

layoutName = "Green"
imagePathPattern = "/tmp/image-%03d.png"
steps = 10

widget = slicer.app.layoutManager().sliceWidget(layoutName)
view = widget.sliceView()
logic = widget.sliceLogic()
bounds = [0,]*6
logic.GetSliceBounds(bounds)

for step in range(steps):
  offset = bounds[4] + step/(1.*steps) * (bounds[5]-bounds[4])
  logic.SetSliceOffset(offset)
  view.forceRender()
  image = qt.QPixmap.grabWidget(view).toImage()
  image.save(imagePathPattern % step)

Segmentations

Load a 3D image or model file as segmentation

slicer.util.loadSegmentation("c:/tmp/tmp/Segmentation.nrrd")
slicer.util.loadSegmentation("c:/tmp/tmp/Segmentation.nii")
slicer.util.loadSegmentation("c:/tmp/Segment_1.stl")

Create a segmentation from a labelmap volume and display in 3D

labelmapVolumeNode = getNode("label")
seg = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLSegmentationNode")
slicer.modules.segmentations.logic().ImportLabelmapToSegmentationNode(labelmapVolumeNode, seg)
seg.CreateClosedSurfaceRepresentation()
slicer.mrmlScene.RemoveNode(labelmapVolumeNode)

The last line is optional. It removes the original labelmap volume so that the same information is not shown twice.

Export labelmap node from segmentation node

Export labelmap matching reference geometry of the segmentation:

segmentationNode = getNode("Segmentation")
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportAllSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode, slicer.vtkSegmentation.EXTENT_REFERENCE_GEOMETRY)

Export smallest possible labelmap:

segmentationNode = getNode("Segmentation")
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportAllSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode)

Export labelmap that matches geometry of a chosen reference volume:

segmentationNode = getNode("Segmentation")
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportVisibleSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode, referenceVolumeNode)

``Export a selection of segments (identified by their names):

segmentNames = ["Prostate", "Urethra"]
segmentIds = vtk.vtkStringArray()
for segmentName in segmentNames:
  segmentId = segmentationNode.GetSegmentation().GetSegmentIdBySegmentName(segmentName)
  segmentIds.InsertNextValue(segmentId)
slicer.vtkSlicerSegmentationsModuleLogic.ExportSegmentsToLabelmapNode(segmentationNode, segmentIds, labelmapVolumeNode, referenceVolumeNode)

Export to file by pressing Ctrl+Shift+S key:

outputPath = "c:/tmp"

def exportLabelmap():
  segmentationNode = slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLSegmentationNode")
  referenceVolumeNode = slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLScalarVolumeNode")
  labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
  slicer.modules.segmentations.logic().ExportVisibleSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode, referenceVolumeNode)
  filepath = outputPath + "/" + referenceVolumeNode.GetName() + "-label.nrrd"
  slicer.util.saveNode(labelmapVolumeNode, filepath)
  slicer.mrmlScene.RemoveNode(labelmapVolumeNode.GetDisplayNode().GetColorNode())
  slicer.mrmlScene.RemoveNode(labelmapVolumeNode)
  slicer.util.delayDisplay("Segmentation saved to " + filepath)

shortcut = qt.QShortcut(slicer.util.mainWindow())
shortcut.setKey(qt.QKeySequence("Ctrl+Shift+s"))
shortcut.connect( "activated()", exportLabelmap)

Export model nodes from segmentation node

segmentationNode = getNode("Segmentation")
shNode = slicer.mrmlScene.GetSubjectHierarchyNode()
exportFolderItemId = shNode.CreateFolderItem(shNode.GetSceneItemID(), "Segments")
slicer.modules.segmentations.logic().ExportAllSegmentsToModels(segmentationNode, exportFolderItemId)

Create a hollow model from boundary of solid segment

In most cases, the most robust and flexible tool for creating empty shell models (e.g., vessel wall model from contrast agent segmentation) is the “Hollow” effect in Segment Editor module. However, for very thin shells, extrusion of the exported surface mesh representation may be just as robust and require less memory and computation time. In this case it may be a better approach to to export the segment to a mesh and extrude it along surface normal direction:

Example using Dynamic Modeler module (allows real-time update of parameters, using GUI in Dynamic Modeler module):

segmentationNode = getNode("Segmentation")

# Export segments to models
shNode = slicer.mrmlScene.GetSubjectHierarchyNode()
exportFolderItemId = shNode.CreateFolderItem(shNode.GetSceneItemID(), "Segments")
slicer.modules.segmentations.logic().ExportAllSegmentsToModels(segmentationNode, exportFolderItemId)
segmentModels = vtk.vtkCollection()
shNode.GetDataNodesInBranch(exportFolderItemId, segmentModels)
# Get exported model of first segment
modelNode = segmentModels.GetItemAsObject(0)

# Set up Hollow tool
hollowModeler = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLDynamicModelerNode")
hollowModeler.SetToolName("Hollow")
hollowModeler.SetNodeReferenceID("Hollow.InputModel", modelNode.GetID())
hollowedModelNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLModelNode")  # this node will store the hollow model
hollowModeler.SetNodeReferenceID("Hollow.OutputModel", hollowedModelNode.GetID())
hollowModeler.SetAttribute("ShellThickness", "2.5")  # grow outside
hollowModeler.SetContinuousUpdate(True)  # auto-update output model if input parameters are changed

# Hide inputs, show output
segmentation.GetDisplayNode().SetVisibility(False)
modelNode.GetDisplayNode().SetVisibility(False)
hollowedModelNode.GetDisplayNode().SetOpacity(0.5)

Example using VTK filters:

# Get closed surface representation of the segment
shellThickness = 3.0  # mm
segmentationNode = getNode("Segmentation")
segmentationNode.CreateClosedSurfaceRepresentation()
polyData = segmentationNode.GetClosedSurfaceInternalRepresentation("Segment_1")

# Create shell
extrude = vtk.vtkLinearExtrusionFilter()
extrude.SetInputData(polyData)
extrude.SetExtrusionTypeToNormalExtrusion()
extrude.SetScaleFactor(shellThickness)

# Compute consistent surface normals
triangle_filter = vtk.vtkTriangleFilter()
triangle_filter.SetInputConnection(extrude.GetOutputPort())
normals = vtk.vtkPolyDataNormals()
normals.SetInputConnection(triangle_filter.GetOutputPort())
normals.FlipNormalsOn()

# Save result into new model node
slicer.modules.models.logic().AddModel(normals.GetOutputPort())

Show a segmentation in 3D

Segmentation can only be shown in 3D if closed surface representation (or other 3D-displayable representation) is available. To create closed surface representation:

segmentation.CreateClosedSurfaceRepresentation()

Get a representation of a segment

Access binary labelmap stored in a segmentation node (without exporting it to a volume node) - if it does not exist, it will return None:

image = slicer.vtkOrientedImageData()
segmentationNode.GetBinaryLabelmapRepresentation(segmentID, image)

Get closed surface, if it does not exist, it will return None:

outputPolyData = vtk.vtkPolyData()
segmentationNode.GetClosedSurfaceRepresentation(segmentID, outputPolyData)

Get binary labelmap representation. If it does not exist then it will be created for that single segment. Applies parent transforms by default (if not desired, another argument needs to be added to the end: false):

import vtkSegmentationCorePython as vtkSegmentationCore
outputOrientedImageData = vtkSegmentationCore.vtkOrientedImageData()
slicer.vtkSlicerSegmentationsModuleLogic.GetSegmentBinaryLabelmapRepresentation(segmentationNode, segmentID, outputOrientedImageData)

Same as above, for closed surface representation:

outputPolyData = vtk.vtkPolyData()
slicer.vtkSlicerSegmentationsModuleLogic.GetSegmentClosedSurfaceRepresentation(segmentationNode, segmentID, outputPolyData)

Convert all segments using default path and conversion parameters

segmentationNode.CreateBinaryLabelmapRepresentation()

Convert all segments using custom path or conversion parameters

Change reference image geometry parameter based on an existing referenceImageData image:

referenceGeometry = slicer.vtkSegmentationConverter.SerializeImageGeometry(referenceImageData)
segmentation.SetConversionParameter(slicer.vtkSegmentationConverter.GetReferenceImageGeometryParameterName(), referenceGeometry)

Re-convert using a modified conversion parameter

Changing smoothing factor for closed surface generation:

import vtkSegmentationCorePython as vtkSegmentationCore
segmentation = getNode("Segmentation").GetSegmentation()

# Turn of surface smoothing
segmentation.SetConversionParameter("Smoothing factor","0.0")

# Recreate representation using modified parameters (and default conversion path)
segmentation.RemoveRepresentation(vtkSegmentationCore.vtkSegmentationConverter.GetSegmentationClosedSurfaceRepresentationName())
segmentation.CreateRepresentation(vtkSegmentationCore.vtkSegmentationConverter.GetSegmentationClosedSurfaceRepresentationName())

Create keyboard shortcut for toggling sphere brush for paint and erase effects

def toggleSphereBrush():
  segmentEditorWidget = slicer.modules.segmenteditor.widgetRepresentation().self().editor
  paintEffect = segmentEditorWidget.effectByName("Paint")
  isSphere = paintEffect.integerParameter("BrushSphere")
  # BrushSphere is "common" parameter (shared between paint and erase)
  paintEffect.setCommonParameter("BrushSphere", 0 if isSphere else 1)

shortcut = qt.QShortcut(slicer.util.mainWindow())
shortcut.setKey(qt.QKeySequence("s"))
shortcut.connect("activated()", toggleSphereBrush)

Customize list of displayed Segment editor effects

Only show Paint and Erase effects:

segmentEditorWidget = slicer.modules.segmenteditor.widgetRepresentation().self().editor
segmentEditorWidget.setEffectNameOrder(["Paint", "Erase"])
segmentEditorWidget.unorderedEffectsVisible = False

Show list of all available effect names:

segmentEditorWidget = slicer.modules.segmenteditor.widgetRepresentation().self().editor
print(segmentEditorWidget.availableEffectNames())

Get centroid of a segment in world (RAS) coordinates

This example shows how to get centroid of a segment in world coordinates and show that position in all slice views.

segmentationNode = getNode("Segmentation")
segmentId = "Segment_1"

# Get array voxel coordinates
import numpy as np
seg=arrayFromSegment(segmentation_node, segmentId)
# numpy array has voxel coordinates in reverse order (KJI instead of IJK)
# and the array is cropped to minimum size in the segmentation
mean_KjiCropped = [coords.mean() for coords in np.nonzero(seg)]

# Get segmentation voxel coordinates
segImage = segmentationNode.GetBinaryLabelmapRepresentation(segmentId)
segImageExtent = segImage.GetExtent()
# origin of the array in voxel coordinates is determined by the start extent
mean_Ijk = [mean_KjiCropped[2], mean_KjiCropped[1], mean_KjiCropped[0]] + np.array([segImageExtent[0], segImageExtent[2], segImageExtent[4]])

# Get segmentation physical coordinates
ijkToWorld = vtk.vtkMatrix4x4()
segImage.GetImageToWorldMatrix(ijkToWorld)
mean_World = [0, 0, 0, 1]
ijkToRas.MultiplyPoint(np.append(mean_Ijk,1.0), mean_World)
mean_World = mean_World[0:3]

# If segmentation node is transformed, apply that transform to get RAS coordinates
transformWorldToRas = vtk.vtkGeneralTransform()
slicer.vtkMRMLTransformNode.GetTransformBetweenNodes(segmentationNode.GetParentTransformNode(), None, transformWorldToRas)
mean_Ras = transformWorldToRas.TransformPoint(mean_World)

# Show mean position value and jump to it in all slice viewers
print(mean_Ras)
slicer.modules.markups.logic().JumpSlicesToLocation(mean_Ras[0], mean_Ras[1], mean_Ras[2], True)

Get histogram of a segmented region

# Generate input data
################################################

# Load master volume
import SampleData
sampleDataLogic = SampleData.SampleDataLogic()
masterVolumeNode = sampleDataLogic.downloadMRBrainTumor1()

# Create segmentation
segmentationNode = slicer.vtkMRMLSegmentationNode()
slicer.mrmlScene.AddNode(segmentationNode)
segmentationNode.CreateDefaultDisplayNodes() # only needed for display
segmentationNode.SetReferenceImageGeometryParameterFromVolumeNode(masterVolumeNode)

# Create segment
tumorSeed = vtk.vtkSphereSource()
tumorSeed.SetCenter(-6, 30, 28)
tumorSeed.SetRadius(25)
tumorSeed.Update()
segmentationNode.AddSegmentFromClosedSurfaceRepresentation(tumorSeed.GetOutput(), "Segment A", [1.0,0.0,0.0])

# Compute histogram
################################################

labelValue = 1  # label value of first segment

# Get segmentation as labelmap volume node
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportVisibleSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode, masterVolumeNode)

# Extract all voxels of the segment as numpy array
volumeArray = slicer.util.arrayFromVolume(masterVolumeNode)
labelArray = slicer.util.arrayFromVolume(labelmapVolumeNode)
segmentVoxels = volumeArray[labelArray==labelValue]

# Compute histogram
import numpy as np
histogram = np.histogram(segmentVoxels, bins=50)

# Plot histogram
################################################

slicer.util.plot(histogram, xColumnIndex = 1)

Get segments visible at a selected position

Show in the console names of segments visible at a markups fiducial position:

segmentationNode = slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLSegmentationNode")
markupsFiducialNode = slicer.mrmlScene.GetFirstNodeByClass("vtkMRMLMarkupsFiducialNode")
sliceViewLabel = "Red"  # any slice view where segmentation node is visible works

def printSegmentNames(unused1=None, unused2=None):

  sliceViewWidget = slicer.app.layoutManager().sliceWidget(sliceViewLabel)
  segmentationsDisplayableManager = sliceViewWidget.sliceView().displayableManagerByClassName("vtkMRMLSegmentationsDisplayableManager2D")
  ras = [0,0,0]
  markupsFiducialNode.GetNthControlPointPositionWorld(0, ras)
  segmentIds = vtk.vtkStringArray()
  segmentationsDisplayableManager.GetVisibleSegmentsForPosition(ras, segmentationNode.GetDisplayNode(), segmentIds)
  for idIndex in range(segmentIds.GetNumberOfValues()):
    segment = segmentationNode.GetSegmentation().GetSegment(segmentIds.GetValue(idIndex))
    print("Segment found at position {0}: {1}".format(ras, segment.GetName()))

# Observe markup node changes
markupsFiducialNode.AddObserver(slicer.vtkMRMLMarkupsPlaneNode.PointModifiedEvent, printSegmentNames)
printSegmentNames()

Set default segmentation options

Allow segments to overlap each other by default:

defaultSegmentEditorNode = slicer.vtkMRMLSegmentEditorNode()
defaultSegmentEditorNode.SetOverwriteMode(slicer.vtkMRMLSegmentEditorNode.OverwriteNone)
slicer.mrmlScene.AddDefaultNode(defaultSegmentEditorNode)

To always make this the default, add the lines above to your .slicerrc.py file.

How to run segment editor effects from a script

Editor effects are complex because they need to handle changing master volumes, undo/redo, masking operations, etc. Therefore, it is recommended to use the effect by instantiating a qMRMLSegmentEditorWidget or use/extract processing logic of the effect and use that from a script.

Use logic of effect from a script

This example shows how to perform operations on segmentations using VTK filters extracted from an effect:

Process segment using a VTK filter

This example shows how to apply a VTK filter to a segment that dilates the image by a specified margin.

segmentationNode = getNode("Segmentation")
segmentId = "Segment_1"
kernelSize = [3,1,5]

# Export segment as vtkImageData (via temporary labelmap volume node)
segmentIds = vtk.vtkStringArray()
segmentIds.InsertNextValue(segmentId)
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportSegmentsToLabelmapNode(segmentationNode, segmentIds, labelmapVolumeNode)

# Process segmentation
segmentImageData = labelmapVolumeNode.GetImageData()
erodeDilate = vtk.vtkImageDilateErode3D()
erodeDilate.SetInputData(segmentImageData)
erodeDilate.SetDilateValue(1)
erodeDilate.SetErodeValue(0)
erodeDilate.SetKernelSize(*kernelSize)
erodeDilate.Update()
segmentImageData.DeepCopy(erodeDilate.GetOutput())

# Import segment from vtkImageData
slicer.modules.segmentations.logic().ImportLabelmapToSegmentationNode(labelmapVolumeNode, segmentationNode, segmentIds)

# Cleanup temporary nodes
slicer.mrmlScene.RemoveNode(labelmapVolumeNode.GetDisplayNode().GetColorNode())
slicer.mrmlScene.RemoveNode(labelmapVolumeNode)

Get information from segmentation nrrd file header

You can use this code snippet to get information from segmentation (.seg.nrrd), for example when creating numpy arrays for generating training data for deep learning networks. This script can be used in any Python environment, not just inside Slicer.

# pip_install("pynrrd")

def read_segmentation_info(filename):
  import nrrd
  header = nrrd.read_header(filename)
  segmentation_info = {}
  segments = []
  segment_index = 0
  while True:
    prefix = "Segment{0}_".format(segment_index)
    if not prefix + "ID" in header.keys():
      break
    segment = {}
    segment["index"] = segment_index
    segment["color"] = [float(i) for i in header[prefix + "Color"].split(" ")]  # Segment0_Color:=0.501961 0.682353 0.501961
    segment["colorAutoGenerated"] = int(header[prefix + "ColorAutoGenerated"]) != 0  # Segment0_ColorAutoGenerated:=1
    segment["extent"] = [int(i) for i in header[prefix + "Extent"].split(" ")]  # Segment0_Extent:=68 203 53 211 24 118
    segment["id"] = header[prefix + "ID"]  # Segment0_ID:=Segment_1
    segment["labelValue"] = int(header[prefix + "LabelValue"])  # Segment0_LabelValue:=1
    segment["layer"] = int(header[prefix + "Layer"])  # Segment0_Layer:=0
    segment["name"] = header[prefix + "Name"]  # Segment0_Name:=Segment_1
    segment["nameAutoGenerated"] = int(header[prefix + "NameAutoGenerated"]) != 0  # Segment0_NameAutoGenerated:=1
    # Segment0_Tags:=Segmentation.Status:inprogress|TerminologyEntry:Segmentation category and type - 3D Slicer General Anatomy list
    # ~SCT^85756007^Tissue~SCT^85756007^Tissue~^^~Anatomic codes - DICOM master list~^^~^^|
    tags = {}
    tags_str = header[prefix + "Tags"].split("|")
    for tag_str in tags_str:
      tag_str = tag_str.strip()
      if not tag_str:
        continue
      key, value = tag_str.split(":", maxsplit=1)
      tags[key] = value
    segment["tags"] = tags
    segments.append(segment)
    segment_index += 1
  segmentation_info["segments"] = segments
  return segmentation_info

def segment_from_name(segmentation_info, segment_name):
  for segment in segmentation_info["segments"]:
    if segment_name == segment["name"]:
      return segment
  raise KeyError("segment not found by name " + segment_name)

def segment_names(segmentation_info):
  names = []
  for segment in segmentation_info["segments"]:
    names.append(segment["name"])
  return names

def extract_segments(voxels, header, segmentation_info, segment_names_to_label_values):
  import numpy as np
  # Create empty array from last 3 dimensions (output will be flattened to a 3D array)
  output_voxels = np.zeros(voxels.shape[-3:])
  # Copy non-segmentation fields to the extracted header
  output_header = {}
  for key in header.keys():
    if not re.match("^Segment[0-9]+_.+", key):
      output_header[key] = header[key]
  # Copy extracted segments
  dims = len(voxels.shape)
  for output_segment_index, segment_name_to_label_value in enumerate(segment_names_to_label_values):
    # Copy relabeled voxel data
    segment = segment_from_name(segmentation_info, segment_name_to_label_value[0])
    input_label_value = segment["labelValue"]
    output_label_value = segment_name_to_label_value[1]
    if dims == 3:
      output_voxels[voxels == input_label_value] = output_label_value
    elif dims == 4:
        inputLayer = segment["layer"]
      output_voxels[voxels[inputLayer,:,:,:] == input_label_value] = output_label_value
    else:
      raise ValueError("Voxel array dimension is invalid")
    # Copy all segment fields corresponding to this segment
    for key in header.keys():
      prefix = "Segment{0}_".format(segment["index"])
      matched = re.match("^" + prefix + "(.+)", key)
      if matched:
        field_name = matched.groups()[0]
        if field_name == "LabelValue":
          value = output_label_value
        elif field_name == "Layer":
          # Output is a single layer (3D volume)
          value = 0
        else:
          value = header[key]
        output_header["Segment{0}_".format(output_segment_index) + field_name] = value
  # Remove unnecessary 4th dimension (volume is collapsed into 3D)
  if dims == 4:
    # Remove "none" from "none (0,1,0) (0,0,-1) (-1.2999954223632812,0,0)"
    output_header["space directions"] = output_header["space directions"][-3:,:]
    # Remove "list" from "list domain domain domain"
    output_header["kinds"] = output_header["kinds"][-3:]
  return output_voxels, output_header

# Read segmentation and show some information about segments
filename = "c:/Users/andra/OneDrive/Projects/SegmentationPynrrd/SegmentationOverlapping.seg.nrrd"
segmentation_info = read_segmentation_info(filename)
number_of_segments = len(segmentation_info["segments"])
names = segment_names(segmentation_info)
label0 = segment_from_name(segmentation_info, names[0])["labelValue"]
print("Number of segments: " + str())
print("Segment names: " + str(names))
print("Label value of {0}: {1}".format(names[0], label0))

# Extract selected segments with chosen label values
extracted_filename = "c:/Users/andra/OneDrive/Projects/SegmentationPynrrd/SegmentationExtracted.seg.nrrd"
voxels, header = nrrd.read(filename)
segment_list = [("Segment_1", 10), ("Segment_3", 12), ("Segment_4", 6)]
extracted_voxels, extracted_header = extract_segments(voxels, header, segmentation_info, segment_list)
nrrd.write(extracted_filename, extracted_voxels, extracted_header)

Quantifying segments

Get centroid of each segment

Place a markups fiducial point at the centroid of each segment.

segmentationNode = getNode("Segmentation")

# Compute centroids
import SegmentStatistics
segStatLogic = SegmentStatistics.SegmentStatisticsLogic()
segStatLogic.getParameterNode().SetParameter("Segmentation", segmentationNode.GetID())
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.centroid_ras.enabled", str(True))
segStatLogic.computeStatistics()
stats = segStatLogic.getStatistics()

# Place a markup point in each centroid
markupsNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLMarkupsFiducialNode")
markupsNode.CreateDefaultDisplayNodes()
for segmentId in stats["SegmentIDs"]:
  centroid_ras = stats[segmentId,"LabelmapSegmentStatisticsPlugin.centroid_ras"]
  segmentName = segmentationNode.GetSegmentation().GetSegment(segmentId).GetName()
  markupsNode.AddFiducialFromArray(centroid_ras, segmentName)

Get size, position, and orientation of each segment

This example computes oriented bounding box for each segment and displays them using annotation ROI.

segmentationNode = getNode("Segmentation")

# Compute bounding boxes
import SegmentStatistics
segStatLogic = SegmentStatistics.SegmentStatisticsLogic()
segStatLogic.getParameterNode().SetParameter("Segmentation", segmentationNode.GetID())
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.obb_origin_ras.enabled",str(True))
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.obb_diameter_mm.enabled",str(True))
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.obb_direction_ras_x.enabled",str(True))
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.obb_direction_ras_y.enabled",str(True))
segStatLogic.getParameterNode().SetParameter("LabelmapSegmentStatisticsPlugin.obb_direction_ras_z.enabled",str(True))
segStatLogic.computeStatistics()
stats = segStatLogic.getStatistics()

# Draw ROI for each oriented bounding box
import numpy as np
for segmentId in stats["SegmentIDs"]:
  # Get bounding box
  obb_origin_ras = np.array(stats[segmentId,"LabelmapSegmentStatisticsPlugin.obb_origin_ras"])
  obb_diameter_mm = np.array(stats[segmentId,"LabelmapSegmentStatisticsPlugin.obb_diameter_mm"])
  obb_direction_ras_x = np.array(stats[segmentId,"LabelmapSegmentStatisticsPlugin.obb_direction_ras_x"])
  obb_direction_ras_y = np.array(stats[segmentId,"LabelmapSegmentStatisticsPlugin.obb_direction_ras_y"])
  obb_direction_ras_z = np.array(stats[segmentId,"LabelmapSegmentStatisticsPlugin.obb_direction_ras_z"])
  # Create ROI
  segment = segmentationNode.GetSegmentation().GetSegment(segmentId)
  roi=slicer.mrmlScene.AddNewNodeByClass("vtkMRMLAnnotationROINode")
  roi.SetName(segment.GetName() + " bounding box")
  roi.SetXYZ(0.0, 0.0, 0.0)
  roi.SetRadiusXYZ(*(0.5*obb_diameter_mm))
  # Position and orient ROI using a transform
  obb_center_ras = obb_origin_ras+0.5*(obb_diameter_mm[0] * obb_direction_ras_x + obb_diameter_mm[1] * obb_direction_ras_y + obb_diameter_mm[2] * obb_direction_ras_z)
  boundingBoxToRasTransform = np.row_stack((np.column_stack((obb_direction_ras_x, obb_direction_ras_y, obb_direction_ras_z, obb_center_ras)), (0, 0, 0, 1)))
  boundingBoxToRasTransformMatrix = slicer.util.vtkMatrixFromArray(boundingBoxToRasTransform)
  transformNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLTransformNode")
  transformNode.SetAndObserveMatrixTransformToParent(boundingBoxToRasTransformMatrix)
  roi.SetAndObserveTransformNodeID(transformNode.GetID())

Complete list of available parameters can be obtained by running segStatLogic.getParameterNode().GetParameterNames().

Sequences

Access voxels of a 4D volume as numpy array

# Get sequence node
import SampleData
sequenceNode = SampleData.SampleDataLogic().downloadSample("CTPCardioSeq")
# Alternatively, get the first sequence node in the scene:
# sequenceNode = slicer.util.getNodesByClass("vtkMRMLSequenceNode")[0]

# Get voxels of itemIndex'th volume as numpy array
itemIndex = 5
voxelArray = slicer.util.arrayFromVolume(sequenceNode.GetNthDataNode(itemIndex))

Get index value

print("Index value of {0}th item: {1} = {2} {3}".format(
  itemIndex,
  sequenceNode.GetIndexName(),
  sequenceNode.GetNthIndexValue(itemIndex),
  sequenceNode.GetIndexUnit()))

Browse a sequence and access currently displayed nodes

# Get a sequence node
import SampleData
sequenceNode = SampleData.SampleDataLogic().downloadSample("CTPCardioSeq")

# Find corresponding sequence browser node
browserNode = slicer.modules.sequences.logic().GetFirstBrowserNodeForSequenceNode(sequenceNode)

# Print sequence information
print("Number of items in the sequence: {0}".format(browserNode.GetNumberOfItems()))
print("Index name: {0}".format(browserNode.GetMasterSequenceNode().GetIndexName()))

# Jump to a selected sequence item
browserNode.SetSelectedItemNumber(5)

# Get currently displayed volume node voxels as numpy array
volumeNode = browserNode.GetProxyNode(sequenceNode)
voxelArray = slicer.util.arrayFromVolume(volumeNode)

Concatenate all sequences in the scene into a new sequence

# Get all sequence nodes in the scene
sequenceNodes = slicer.util.getNodesByClass("vtkMRMLSequenceNode")
mergedSequenceNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLSequenceNode", "Merged sequence")

# Merge all sequence nodes into a new sequence node
mergedIndexValue = 0
for sequenceNode in sequenceNodes:
  for itemIndex in range(sequenceNode.GetNumberOfDataNodes()):
    dataNode = sequenceNode.GetNthDataNode(itemIndex)
    mergedSequenceNode.SetDataNodeAtValue(dataNode, str(mergedIndexValue))
    mergedIndexValue += 1
  # Delete the sequence node we copied the data from, to prevent sharing of the same
  # node by multiple sequences
  slicer.mrmlScene.RemoveNode(sequenceNode)

# Create a sequence browser node for the new merged sequence
mergedSequenceBrowserNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLSequenceBrowserNode", "Merged")
mergedSequenceBrowserNode.AddSynchronizedSequenceNode(mergedSequenceNode)
slicer.modules.sequencebrowser.setToolBarActiveBrowserNode(mergedSequenceBrowserNode)
# Show proxy node in slice viewers
mergedProxyNode = mergedSequenceBrowserNode.GetProxyNode(mergedSequenceNode)
slicer.util.setSliceViewerLayers(background=mergedProxyNode)

Subject hierarchy

Get the pseudo-singleton subject hierarchy node

It manages the whole hierarchy and provides functions to access and manipulate

shNode = slicer.mrmlScene.GetSubjectHierarchyNode()

Create subject hierarchy item

# If it is for a data node, it is automatically created, but the create function can be used to set parent:
shNode.CreateItem(parentItemID, dataNode)
# If it is a hierarchy item without a data node, then the create function must be used:
shNode.CreateSubjectItem(parentItemID, name)
shNode.CreateFolderItem(parentItemID, name)
shNode.CreateHierarchyItem(parentItemID, name, level) # Advanced method to set level attribute manually (usually subject, study, or folder, but it can be a virtual branch for example)

Get subject hierarchy item

Items in subject hierarchy are uniquely identified by integer IDs

# Get scene item ID first because it is the root item:
sceneItemID = shNode.GetSceneItemID()
# Get direct child by name
subjectItemID = shNode.GetItemChildWithName(sceneItemID, "Subject_1")
# Get item for data node
itemID = shNode.GetItemByDataNode(dataNode)
# Get item by UID (such as DICOM)
itemID = shNode.GetItemByUID(slicer.vtkMRMLSubjectHierarchyConstants.GetDICOMUIDName(), seriesInstanceUid)
itemID = shNode.GetItemByUIDList(slicer.vtkMRMLSubjectHierarchyConstants.GetDICOMInstanceUIDName(), instanceUID)
# Invalid item ID for checking validity of a given ID (most functions return the invalid ID when item is not found)
invalidItemID = slicer.vtkMRMLSubjectHierarchyNode.GetInvalidItemID()

Traverse children of a subject hierarchy item

children = vtk.vtkIdList()
shNode.GetItemChildren(parent, children) # Add a third argument with value True for recursive query
for i in range(children.GetNumberOfIds()):
  child = children.GetId(i)
  ...

Manipulate subject hierarchy item

Instead of node operations on the individual subject hierarchy nodes, item operations are performed on the one subject hierarchy node.

# Set item name
shNode.SetItemName(itemID, "NewName")
# Set item parent (reparent)
shNode.SetItemParent(itemID, newParentItemID)
# Set visibility of data node associated to an item
shNode.SetItemDisplayVisibility(itemID, 1)
# Set visibility of whole branch
# Note: Folder-type items (fodler, subject, study, etc.) create their own display nodes when show/hiding from UI.
#       The displayable managers use SH information to determine visibility of an item, so no need to show/hide individual leaf nodes any more.
#       Once the folder display node is created, it can be shown hidden simply using shNode.SetItemDisplayVisibility
# From python, this is how to trigger creating a folder display node
pluginHandler = slicer.qSlicerSubjectHierarchyPluginHandler().instance()
folderPlugin = pluginHandler.pluginByName("Folder")
folderPlugin.setDisplayVisibility(folderItemID, 1)

Filter items in TreeView or ComboBox

Displayed items can be filtered using setAttributeFilter method. An example of the usage can be found in the unit test. Modified version here:

print(shTreeView.displayedItemCount()) # 5
shTreeView.setAttributeFilter("DICOM.Modality") # Nodes must have this attribute
print(shTreeView.displayedItemCount()) # 3
shTreeView.setAttributeFilter("DICOM.Modality","CT") # Have attribute and equal ``CT``
print(shTreeView.displayedItemCount()) # 1
shTreeView.removeAttributeFilter()
print(shTreeView.displayedItemCount()) # 5

Listen to subject hierarchy item events

The subject hierarchy node sends the node item id as calldata. Item IDs are vtkIdType, which are NOT vtkObjects. You need to use vtk.calldata_type(vtk.VTK_LONG) (otherwise the application crashes).

class MyListenerClass(VTKObservationMixin):
  def __init__(self):
    VTKObservationMixin.__init__(self)

    shNode = slicer.vtkMRMLSubjectHierarchyNode.GetSubjectHierarchyNode(slicer.mrmlScene)
    self.addObserver(shNode, shNode.SubjectHierarchyItemModifiedEvent, self.shItemModifiedEvent)

  @vtk.calldata_type(vtk.VTK_LONG)
  def shItemModifiedEvent(self, caller, eventId, callData):
    print("SH Node modified")
    print("SH item ID: {0}".format(callData))

Subject hierarchy plugin offering view context menu action

If an object that supports view context menus (e.g. markups) is right-clicked in a slice or 3D view, it can offer custom actions. Due to internal limitations these plugins must be set up differently, as explained here. This example makes it easier to create such a plugin.

import vtk, qt, ctk, slicer
from slicer.ScriptedLoadableModule import *
from slicer.util import VTKObservationMixin

from SubjectHierarchyPlugins import AbstractScriptedSubjectHierarchyPlugin

class ViewContextMenu(ScriptedLoadableModule):
"""Uses ScriptedLoadableModule base class, available at:
  https://github.com/Slicer/Slicer/blob/master/Base/Python/slicer/ScriptedLoadableModule.py
  """

  def __init__(self, parent):
    ScriptedLoadableModule.__init__(self, parent)
    self.parent.title = "Markup Editor"
    self.parent.categories = ["SlicerMorph", "Labs"]
    self.parent.dependencies = []
    self.parent.contributors = ["Steve Pieper (Isomics, Inc.)"]
    self.parent.helpText = """
A tool to manipulate Markups using the Segment Editor as a geometry backend
"""
    self.parent.helpText += self.getDefaultModuleDocumentationLink()
    self.parent.acknowledgementText = """
This module was developed by Steve Pieper, Sara Rolfe and Murat Maga,
through a NSF ABI Development grant, "An Integrated Platform for Retrieval,
Visualization and Analysis of 3D Morphology From Digital Biological Collections"
(Award Numbers: 1759883 (Murat Maga), 1759637 (Adam Summers), 1759839 (Douglas Boyer)).
This file was originally developed by Jean-Christophe Fillion-Robin, Kitware Inc.,
Andras Lasso, PerkLab, and Steve Pieper, Isomics, Inc.
and was partially funded by NIH grant 3P41RR013218-12S1.
"""

    #
    # register subject hierarchy plugin once app is initialized
    #
    def onStartupCompleted():
      import SubjectHierarchyPlugins
      from ViewContextMenu import ViewContextMenuSubjectHierarchyPlugin
      scriptedPlugin = slicer.qSlicerSubjectHierarchyScriptedPlugin(None)
      scriptedPlugin.setPythonSource(ViewContextMenuSubjectHierarchyPlugin.filePath)
      pluginHandler = slicer.qSlicerSubjectHierarchyPluginHandler.instance()
      pluginHandler.registerPlugin(scriptedPlugin)
      print("ViewContextMenuSubjectHierarchyPlugin loaded")

    slicer.app.connect("startupCompleted()", onStartupCompleted)


class ViewContextMenuSubjectHierarchyPlugin(AbstractScriptedSubjectHierarchyPlugin):

  # Necessary static member to be able to set python source to scripted subject hierarchy plugin
  filePath = __file__

  def __init__(self, scriptedPlugin):
    self.viewAction = qt.QAction(f"CUSTOM VIEW ...", scriptedPlugin)
    self.viewAction.objectName = "CustomViewAction"
    self.viewAction.connect("triggered()", self.onViewAction)

  def onViewAction(self):
    print(f"VIEW ACTION")

  def viewContextMenuActions(self):
    return [self.viewAction]

  def showViewContextMenuActionsForItem(self, itemID, eventData=None):
    pluginHandler = slicer.qSlicerSubjectHierarchyPluginHandler.instance()
    pluginLogic = pluginHandler.pluginLogic()
    menuActions = list(pluginLogic.availableViewMenuActionNames())
    menuActions.append("CustomViewAction")
    pluginLogic.setDisplayedViewMenuActionNames(menuActions)
      self.viewAction.visible = True

Use whitelist to customize view menu

When right-clicking certain types of nodes in the 2D/3D views, a subject hierarchy menu pops up. If menu actions need to be removed, a whitelist can be used to specify the ones that should show up.

pluginHandler = slicer.qSlicerSubjectHierarchyPluginHandler.instance()
pluginLogic = pluginHandler.pluginLogic()
menuActions = pluginLogic.availableViewMenuActionNames()
# Returns ("RenamePointAction", "DeletePointAction", "ToggleSelectPointAction", "EditPropertiesAction")
newActions = ["RenamePointAction"]
pluginLogic.setDisplayedViewMenuActionNames(newActions)

Tractography

Export a tract (FiberBundle) to Blender, including color

Note: an interactive version of this script is now included in the SlicerDMRI extension (module code). After installing SlicerDMRI, go to Modules -> Diffusion -> Import and Export -> Export tractography to PLY (mesh).

The example below shows how to export a tractography “FiberBundleNode” to a PLY file:

lineDisplayNode = getNode("*LineDisplay*")
plyFilePath = "/tmp/fibers.ply"

tuber = vtk.vtkTubeFilter()
tuber.SetInputData(lineDisplayNode.GetOutputPolyData())
tuber.Update()
tubes = tuber.GetOutputDataObject(0)
scalars = tubes.GetPointData().GetArray(0)
scalars.SetName("scalars")

triangles = vtk.vtkTriangleFilter()
triangles.SetInputData(tubes)
triangles.Update()

colorNode = lineDisplayNode.GetColorNode()
lookupTable = vtk.vtkLookupTable()
lookupTable.DeepCopy(colorNode.GetLookupTable())
lookupTable.SetTableRange(0,1)

plyWriter = vtk.vtkPLYWriter()
plyWriter.SetInputData(triangles.GetOutput())
plyWriter.SetLookupTable(lookupTable)
plyWriter.SetArrayName("scalars")

plyWriter.SetFileName(plyFilePath)
plyWriter.Write()

Iterate over tract (FiberBundle) streamline points

This example shows how to access the points in each line of a FiberBundle as a numpy array (view).

from vtk.util.numpy_support import vtk_to_numpy

fb = getNode("FiberBundle_F") # <- fill in node ID here

# get point data as 1d array
points = slicer.util.arrayFromModelPoints(fb)

# get line cell ids as 1d array
line_ids = vtk_to_numpy(fb.GetPolyData().GetLines().GetData())

# VTK cell ids are stored as
#   [ N0 c0_id0 ... c0_id0
#     N1 c1_id0 ... c1_idN1 ]
# so we need to
# - read point count for each line (cell)
# - grab the ids in that range from `line_ids` array defined above
# - index the `points` array by those ids
cur_idx = 1
for _ in range(pd.GetLines().GetNumberOfCells()):
  # - read point count for this line (cell)
  count = lines[cur_idx - 1]

  # - grab the ids in that range from `lines`
  index_array = line_ids[ cur_idx : cur_idx + count]
  # update to the next range
  cur_idx += count + 1

  # - index the point array by those ids
  line_points = points[index_array]

  # do work here

Transforms

Get a notification if a transform is modified

def onTransformNodeModified(transformNode, unusedArg2=None, unusedArg3=None):
  transformMatrix = vtk.vtkMatrix4x4()
  transformNode.GetMatrixTransformToWorld(transformMatrix)
  print("Position: [{0}, {1}, {2}]".format(transformMatrix.GetElement(0,3), transformMatrix.GetElement(1,3), transformMatrix.GetElement(2,3)))

transformNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLTransformNode")
transformNode.AddObserver(slicer.vtkMRMLTransformNode.TransformModifiedEvent, onTransformNodeModified)

Rotate a node around a specified point

Set up the scene:

  • Add a markup fiducial node (centerOfRotationMarkupsNode) with a single point to specify center of rotation.
  • Add a rotation transform (rotationTransformNode) that will be edited in Transforms module to specify rotation angles.
  • Add a transform (finalTransformNode) and apply it (not harden) to those nodes (images, models, etc.) that you want to rotate around the center of rotation point.

Then run the script below, go to Transforms module, select rotationTransformNode, and move rotation sliders.

# This markups fiducial node specifies the center of rotation
centerOfRotationMarkupsNode = getNode("F")
# This transform can be  edited in Transforms module
rotationTransformNode = getNode("LinearTransform_3")
# This transform has to be applied to the image, model, etc.
finalTransformNode = getNode("LinearTransform_4")

def updateFinalTransform(unusedArg1=None, unusedArg2=None, unusedArg3=None):
  rotationMatrix = vtk.vtkMatrix4x4()
  rotationTransformNode.GetMatrixTransformToParent(rotationMatrix)
  rotationCenterPointCoord = [0.0, 0.0, 0.0]
  centerOfRotationMarkupsNode.GetNthControlPointPositionWorld(0, rotationCenterPointCoord)
  finalTransform = vtk.vtkTransform()
  finalTransform.Translate(rotationCenterPointCoord)
  finalTransform.Concatenate(rotationMatrix)
  finalTransform.Translate(-rotationCenterPointCoord[0], -rotationCenterPointCoord[1], -rotationCenterPointCoord[2])
  finalTransformNode.SetAndObserveMatrixTransformToParent(finalTransform.GetMatrix())

# Manual initial update
updateFinalTransform()

# Automatic update when point is moved or transform is modified
rotationTransformNodeObserver = rotationTransformNode.AddObserver(slicer.vtkMRMLTransformNode.TransformModifiedEvent, updateFinalTransform)
centerOfRotationMarkupsNodeObserver = centerOfRotationMarkupsNode.AddObserver(slicer.vtkMRMLMarkupsNode.PointModifiedEvent, updateFinalTransform)

# Execute these lines to stop automatic updates:
# rotationTransformNode.RemoveObserver(rotationTransformNodeObserver)
# centerOfRotationMarkupsNode.RemoveObserver(centerOfRotationMarkupsNodeObserver)

Rotate a node around a specified line

Set up the scene:

  • Add a markup line node (rotationAxisMarkupsNode) with 2 points to specify rotation axis.
  • Add a rotation transform (rotationTransformNode) that will be edited in Transforms module to specify rotation angle.
  • Add a transform (finalTransformNode) and apply it (not harden) to those nodes (images, models, etc.) that you want to rotate around the line.

Then run the script below, go to Transforms module, select rotationTransformNode, and move Edit / Rotation / IS slider.

# This markups fiducial node specifies the center of rotation
rotationAxisMarkupsNode = getNode("L")
# This transform can be edited in Transforms module (Edit / Rotation / IS slider)
rotationTransformNode = getNode("LinearTransform_3")
# This transform has to be applied to the image, model, etc.
finalTransformNode = getNode("LinearTransform_4")

def updateFinalTransform(unusedArg1=None, unusedArg2=None, unusedArg3=None):
  import numpy as np
  rotationAxisPoint1_World = np.zeros(3)
  rotationAxisMarkupsNode.GetNthControlPointPositionWorld(0, rotationAxisPoint1_World)
  rotationAxisPoint2_World = np.zeros(3)
  rotationAxisMarkupsNode.GetNthControlPointPositionWorld(1, rotationAxisPoint2_World)
  axisDirectionZ_World = rotationAxisPoint2_World-rotationAxisPoint1_World
  axisDirectionZ_World = axisDirectionZ_World/np.linalg.norm(axisDirectionZ_World)
  # Get transformation between world coordinate system and rotation axis aligned coordinate system
  worldToRotationAxisTransform = vtk.vtkMatrix4x4()
  p=vtk.vtkPlaneSource()
  p.SetNormal(axisDirectionZ_World)
  axisOrigin = np.array(p.GetOrigin())
  axisDirectionX_World = np.array(p.GetPoint1())-axisOrigin
  axisDirectionY_World = np.array(p.GetPoint2())-axisOrigin
  rotationAxisToWorldTransform = np.row_stack((np.column_stack((axisDirectionX_World, axisDirectionY_World, axisDirectionZ_World, rotationAxisPoint1_World)), (0, 0, 0, 1)))
  rotationAxisToWorldTransformMatrix = slicer.util.vtkMatrixFromArray(rotationAxisToWorldTransform)
  worldToRotationAxisTransformMatrix = slicer.util.vtkMatrixFromArray(np.linalg.inv(rotationAxisToWorldTransform))
  # Compute transformation chain
  rotationMatrix = vtk.vtkMatrix4x4()
  rotationTransformNode.GetMatrixTransformToParent(rotationMatrix)
  finalTransform = vtk.vtkTransform()
  finalTransform.Concatenate(rotationAxisToWorldTransformMatrix)
  finalTransform.Concatenate(rotationMatrix)
  finalTransform.Concatenate(worldToRotationAxisTransformMatrix)
  finalTransformNode.SetAndObserveMatrixTransformToParent(finalTransform.GetMatrix())

# Manual initial update
updateFinalTransform()

# Automatic update when point is moved or transform is modified
rotationTransformNodeObserver = rotationTransformNode.AddObserver(slicer.vtkMRMLTransformNode.TransformModifiedEvent, updateFinalTransform)
rotationAxisMarkupsNodeObserver = rotationAxisMarkupsNode.AddObserver(slicer.vtkMRMLMarkupsNode.PointModifiedEvent, updateFinalTransform)

# Execute these lines to stop automatic updates:
# rotationTransformNode.RemoveObserver(rotationTransformNodeObserver)
# rotationAxisMarkupsNode.RemoveObserver(rotationAxisMarkupsNodeObserver)

Convert between ITK and Slicer linear transforms

# Copy the content between the following triple-quotes to a file called 'LinearTransform.tfm', and load into Slicer

tfm_file = """#Insight Transform File V1.0
#Transform 0
Transform: AffineTransform_double_3_3
Parameters: 0.929794207512361 0.03834792453582355 -0.3660767246906854 -0.2694570325150706 0.7484457003494506 -0.6059884002657121 0.2507501531497781 0.6620864522947292 0.7062335947709847 -46.99999999999999 49 17.00000000000002
FixedParameters: 0 0 0"""

import numpy as np

# get the upper 3x4 transform matrix
m = np.array( tfm_file.splitlines()[3].split()[1:], dtype=np.float64 )

# pad to a 4x4 matrix
m2 = np.vstack((m.reshape(4,3).T, [0,0,0,1]))

def itktfm_to_slicer(tfm):
  ras2lps = np.diag([-1, -1, 1, 1])
  mt = ras2lps @ m2 @ ras2lps
  mt[:3,3] = mt[:3,:3] @ mt[:3,3]
  return mt

print( itktfm_to_slicer(m2) )

# Running the code above in Python should print the following output.
# This output should match the display the loaded .tfm file in the Transforms module:
# [[  0.92979  -0.26946  -0.25075  52.64097]
# [  0.03835   0.74845  -0.66209 -46.12696]
# [  0.36608   0.60599   0.70623  -0.48185]
# [  0.        0.        0.        1.     ]]

C++:

// Convert from LPS (ITK) to RAS (Slicer)
// input: transformVtk_LPS matrix in vtkMatrix4x4 in resampling convention in LPS
// output: transformVtk_RAS matrix in vtkMatri4x4 in modeling convention in RAS

// Tras = lps2ras * Tlps * ras2lps
vtkSmartPointer<vtkMatrix4x4> lps2ras = vtkSmartPointer<vtkMatrix4x4>::New();
lps2ras->SetElement(0,0,-1);
lps2ras->SetElement(1,1,-1);
vtkMatrix4x4* ras2lps = lps2ras; // lps2ras is diagonal therefore the inverse is identical
vtkMatrix4x4::Multiply4x4(lps2ras, transformVtk_LPS, transformVtk_LPS);
vtkMatrix4x4::Multiply4x4(transformVtk_LPS, ras2lps, transformVtk_RAS);

// Convert the sense of the transform (from ITK resampling to Slicer modeling transform)
vtkMatrix4x4::Invert(transformVtk_RAS);

Apply a transform to a transformable node

transformToParentMatrix = vtk.vtkMatrix4x4()
...
transformNode.SetMatrixTransformToParent(matrix)
transformableNode.SetAndObserveTransformNodeID(transformNode.GetID())

C++:

vtkNew<vtkMRMLTransformNode> transformNode;
scene->AddNode(transformNode.GetPointer());
...
vtkNew<vtkMatrix4x4> matrix;
...
transform->SetMatrixTransformToParent( matrix.GetPointer() );
...
vtkMRMLVolumeNode* transformableNode = ...; // or vtkMRMLModelNode*...
transformableNode->SetAndObserveTransformNodeID( transformNode->GetID() );

Set a transformation matrix from a numpy array

# Create a 4x4 transformation matrix as numpy array
transformNode = ...
transformMatrixNP = np.array(
  [[0.92979,-0.26946,-0.25075,52.64097],
  [0.03835, 0.74845, -0.66209, -46.12696],
  [0.36608, 0.60599, 0.70623, -0.48185],
  [0, 0, 0, 1]])

# Update matrix in transform node
transformNode.SetAndObserveMatrixTransformToParent(slicer.util.vtkMatrixFromArray(transformMatrixNP))

Example of moving a volume along a trajectory using a transform

# Load sample volume
import SampleData
sampleDataLogic = SampleData.SampleDataLogic()
mrHead = sampleDataLogic.downloadMRHead()

# Create transform and apply to sample volume
transformNode = slicer.vtkMRMLTransformNode()
slicer.mrmlScene.AddNode(transformNode)
mrHead.SetAndObserveTransformNodeID(transformNode.GetID())

# How to move a volume along a trajectory using a transform:
import time
import math
transformMatrix = vtk.vtkMatrix4x4()
for xPos in range(-30,30):
  transformMatrix.SetElement(0,3, xPos)
  transformMatrix.SetElement(1,3, math.sin(xPos)*10)
  transformNode.SetMatrixTransformToParent(transformMatrix)
  slicer.app.processEvents()
  time.sleep(0.02)
# Note: for longer animations use qt.QTimer.singleShot(100, callbackFunction)
# instead of a for loop.

Combine multiple transforms

Because a transform node is also a transformable node, it is possible to concatenate transforms with each other.

transformNode2.SetAndObserveTransformNodeID(transformNode1.GetID())
transformableNode.SetAndObserveTransformNodeID(transformNode2.GetID())

C++:

vtkMRMLTransformNode* transformNode1 = ...;
vtkMRMLTransformNode* transformNode2 = ...;
transformNode2->SetAndObserveTransformNodeID(transformNode1->GetID());
transformable->SetAndObserveTransformNodeID(transformNode2->GetID());

Convert the transform to a grid transform

Any transform can be converted to a grid transform (also known as displacement field transform):

transformNode=slicer.util.getNode('LinearTransform_3')
referenceVolumeNode=slicer.util.getNode('MRHead')
slicer.modules.transforms.logic().ConvertToGridTransform(transformNode, referenceVolumeNode)

Note

  • Conversion to grid transform is useful because some software cannot use inverse transforms or can only use grid transforms.
  • Displacement field transforms are saved to file differently than displacement field volumes: displacement vectors in transforms are converted to LPS coordinate system on saving, displacement vectors in volumes are saved to file unchanged.

Export the displacement magnitude of the transform as a volume

transformNode=slicer.util.getNode('LinearTransform_3')
referenceVolumeNode=slicer.util.getNode('MRHead')
slicer.modules.transforms.logic().CreateDisplacementVolumeFromTransform(transformNode, referenceVolumeNode, False)

Visualize the displacement magnitude as a color volume

transformNode=slicer.util.getNode('LinearTransform_3')
referenceVolumeNode=slicer.util.getNode('MRHead')
slicer.modules.transforms.logic().CreateDisplacementVolumeFromTransform(transformNode, referenceVolumeNode, True)

Volumes

Load volume from file

loadedVolumeNode = slicer.util.loadVolume("c:/Users/abc/Documents/MRHead.nrrd")

Additional options may be specified in properties argument. For example, load an image stack by disabling singleFile option:

loadedVolumeNode = slicer.util.loadVolume("c:/Users/abc/Documents/SomeImage/file001.png", {"singleFile": False})

Note

The following options can be passed to load volumes programmatically when using qSlicerVolumesReader: - name (string): Node name to set for the loaded volume - labelmap (bool, default=false): Load the file as labelmap volume - singleFile (bool, default=false): Force loading this file only (otherwise the loader may look for similar files in the same folder to load multiple slices as a 3D volume) - autoWindowLevel (bool, default=true): Automatically compute the window level based on the volume pixel intensities - show (bool, default=true): Show the volume in views after loading - center (bool, default=false): Apply a transform that places the volume in the patient coordinate system origin - discardOrientation (bool, default=false): Discard file orientation information. - fileNames (string list): List of files to be loaded as a volume - colorNodeID (string): ID of the color node used to display the volume. Default is vtkMRMLColorTableNodeGrey for scalar volume and vtkMRMLColorTableNodeFileGenericColors.txt for labelmap volume.

Load volume from .vti file

Slicer does not provide reader for VTK XML image data file format (as they are not commonly used for storing medical images and they cannot store image axis directions) but such files can be read by using this script:

reader=vtk.vtkXMLImageDataReader()
reader.SetFileName("/path/to/file.vti")
reader.Update()
imageData = reader.GetOutput()
spacing = imageData.GetSpacing()
origin = imageData.GetOrigin()
imageData.SetOrigin(0,0,0)
imageData.SetSpacing(1,1,1)
volumeNode=slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode")
volumeNode.SetAndObserveImageData(imageData)
volumeNode.SetSpacing(spacing)
volumeNode.SetOrigin(origin)
slicer.util.setSliceViewerLayers(volumeNode, fit=True)

Load volume from a remote server

Download a volume from a remote server by an URL and load it into the scene using the code snippets below.

Note

Downloaded data is temporarily stored in the application’s cache folder and if the checksum of the already downloaded data matches the specified checksum (:) then the file is retrieved from the cache instead of being downloaded again. To compute digest with algo SHA256, you can run slicer.util.computeChecksum("SHA256", "path/to/file")().

Simple download

import SampleData
sampleDataLogic = SampleData.SampleDataLogic()
loadedNodes = sampleDataLogic.downloadFromURL(
  nodeNames="MRHead",
  fileNames="MR-head25.nrrd",
  uris="https://github.com/Slicer/SlicerTestingData/releases/download/SHA256/cc211f0dfd9a05ca3841ce1141b292898b2dd2d3f08286affadf823a7e58df93",
  checksums="SHA256:cc211f0dfd9a05ca3841ce1141b292898b2dd2d3f08286affadf823a7e58df93")[0]

Download with interruptible progress reporting

import SampleData

def reportProgress(msg, level=None):
  # Print progress in the console
  print("Loading... {0}%".format(sampleDataLogic.downloadPercent))
  # Abort download if cancel is clicked in progress bar
  if slicer.progressWindow.wasCanceled:
    raise Exception("download aborted")
  # Update progress window
  slicer.progressWindow.show()
  slicer.progressWindow.activateWindow()
  slicer.progressWindow.setValue(int(sampleDataLogic.downloadPercent))
  slicer.progressWindow.setLabelText("Downloading...")
  # Process events to allow screen to refresh
  slicer.app.processEvents()

try:
  volumeNode = None
  slicer.progressWindow = slicer.util.createProgressDialog()
  sampleDataLogic = SampleData.SampleDataLogic()
  sampleDataLogic.logMessage = reportProgress
  loadedNodes = sampleDataLogic.downloadFromURL(
    nodeNames="MRHead",
    fileNames="MR-head25.nrrd",
    uris="https://github.com/Slicer/SlicerTestingData/releases/download/SHA256/cc211f0dfd9a05ca3841ce1141b292898b2dd2d3f08286affadf823a7e58df93",
    checksums="SHA256:cc211f0dfd9a05ca3841ce1141b292898b2dd2d3f08286affadf823a7e58df93")
  volumeNode = loadedNodes[0]
finally:
  slicer.progressWindow.close()

Show volume rendering automatically when a volume is loaded

To show volume rendering of a volume automatically when it is loaded, add the lines below to your .slicerrc.py file.

@vtk.calldata_type(vtk.VTK_OBJECT)
def onNodeAdded(caller, event, calldata):
  node = calldata
  if isinstance(node, slicer.vtkMRMLVolumeNode):
    # Call showVolumeRendering using a timer instead of calling it directly
    # to allow the volume loading to fully complete.
    qt.QTimer.singleShot(0, lambda: showVolumeRendering(node))

def showVolumeRendering(volumeNode):
  print("Show volume rendering of node " + volumeNode.GetName())
  volRenLogic = slicer.modules.volumerendering.logic()
  displayNode = volRenLogic.CreateDefaultVolumeRenderingNodes(volumeNode)
  displayNode.SetVisibility(True)
  scalarRange = volumeNode.GetImageData().GetScalarRange()
  if scalarRange[1]-scalarRange[0] < 1500:
    # Small dynamic range, probably MRI
    displayNode.GetVolumePropertyNode().Copy(volRenLogic.GetPresetByName("MR-Default"))
  else:
    # Larger dynamic range, probably CT
    displayNode.GetVolumePropertyNode().Copy(volRenLogic.GetPresetByName("CT-Chest-Contrast-Enhanced"))

slicer.mrmlScene.AddObserver(slicer.vtkMRMLScene.NodeAddedEvent, onNodeAdded)

Automatically load volumes that are copied into a folder

This example shows how to implement a simple background task by using a timer. The background task is to check for any new volume files in folder and if there is any then automatically load it.

There are more efficient methods for file system monitoring or exchanging image data in real-time (for example, using OpenIGTLink), the example below is just for demonstration purposes.

incomingVolumeFolder = "c:/tmp/incoming"
incomingVolumesProcessed = []

def checkForNewVolumes():
  # Check if there is a new file in the
  from os import listdir
  from os.path import isfile, join
  for f in listdir(incomingVolumeFolder):
    if f in incomingVolumesProcessed:
      # This is an incoming file, it was already there
      continue
    filePath = join(incomingVolumeFolder, f)
    if not isfile(filePath):
      # ignore directories
      continue
    logging.info("Loading new file: " + f)
    incomingVolumesProcessed.append(f)
    slicer.util.loadVolume(filePath)
  # Check again in 3000ms
  qt.QTimer.singleShot(3000, checkForNewVolumes)

# Start monitoring
checkForNewVolumes()

Extract randomly oriented slabs of given shape from a volume

Returns a numpy array of sliceCount random tiles.

def randomSlices(volume, sliceCount, sliceShape):
  layoutManager = slicer.app.layoutManager()
  redWidget = layoutManager.sliceWidget("Red")
  sliceNode = redWidget.mrmlSliceNode()
  sliceNode.SetDimensions(*sliceShape, 1)
  sliceNode.SetFieldOfView(*sliceShape, 1)
  bounds = [0]*6
  volume.GetRASBounds(bounds)
  imageReslice = redWidget.sliceLogic().GetBackgroundLayer().GetReslice()

  sliceSize = sliceShape[0] * sliceShape[1]
  X = numpy.zeros([sliceCount, sliceSize])

  for sliceIndex in range(sliceCount):
    position = numpy.random.rand(3) * 2 - 1
    position = [bounds[0] + bounds[1]-bounds[0] * position[0],
                bounds[2] + bounds[3]-bounds[2] * position[1],
                bounds[4] + bounds[5]-bounds[4] * position[2]]
    normal = numpy.random.rand(3) * 2 - 1
    normal = normal / numpy.linalg.norm(normal)
    transverse = numpy.cross(normal, [0,0,1])
    orientation = 0
    sliceNode.SetSliceToRASByNTP( normal[0], normal[1], normal[2],
                                  transverse[0], transverse[1], transverse[2],
                                  position[0], position[1], position[2],
                                  orientation)
    if sliceIndex % 100 == 0:
      slicer.app.processEvents()
    imageReslice.Update()
    imageData = imageReslice.GetOutputDataObject(0)
    array = vtk.util.numpy_support.vtk_to_numpy(imageData.GetPointData().GetScalars())
    X[sliceIndex] = array
  return X

Clone a volume

This example shows how to clone the MRHead sample volume, including its pixel data and display settings.

sourceVolumeNode = slicer.util.getNode("MRHead")
volumesLogic = slicer.modules.volumes.logic()
clonedVolumeNode = volumesLogic.CloneVolume(slicer.mrmlScene, sourceVolumeNode, "Cloned volume")

Create a new volume

This example shows how to create a new empty volume. The “Image Maker” extension contains a module that allows creating a volume from scratch without programming.

nodeName = "MyNewVolume"
imageSize = [512, 512, 512]
voxelType=vtk.VTK_UNSIGNED_CHAR
imageOrigin = [0.0, 0.0, 0.0]
imageSpacing = [1.0, 1.0, 1.0]
imageDirections = [[1,0,0], [0,1,0], [0,0,1]]
fillVoxelValue = 0

# Create an empty image volume, filled with fillVoxelValue
imageData = vtk.vtkImageData()
imageData.SetDimensions(imageSize)
imageData.AllocateScalars(voxelType, 1)
imageData.GetPointData().GetScalars().Fill(fillVoxelValue)
# Create volume node
volumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode", nodeName)
volumeNode.SetOrigin(imageOrigin)
volumeNode.SetSpacing(imageSpacing)
volumeNode.SetIJKToRASDirections(imageDirections)
volumeNode.SetAndObserveImageData(imageData)
volumeNode.CreateDefaultDisplayNodes()
volumeNode.CreateDefaultStorageNode()

C++:

vtkNew<vtkImageData> imageData;
imageData->SetDimensions(10,10,10); // image size
imageData->AllocateScalars(VTK_UNSIGNED_CHAR, 1); // image type and number of components
// initialize the pixels here

vtkNew<vtkMRMLScalarVolumeNode> volumeNode;
volumeNode->SetAndObserveImageData(imageData);
volumeNode->SetOrigin( -10., -10., -10.);
volumeNode->SetSpacing( 2., 2., 2. );
mrmlScene->AddNode( volumeNode.GetPointer() );

volumeNode->CreateDefaultDisplayNodes()

Note

Origin and spacing must be set on the volume node instead of the image data.

Get value of a volume at specific voxel coordinates

This example shows how to get voxel value of “volumeNode” at “ijk” volume voxel coordinates.

volumeNode = slicer.util.getNode("MRHead")
ijk = [20,40,30]  # volume voxel coordinates

voxels = slicer.util.arrayFromVolume(volumeNode)  # get voxels as a numpy array
voxelValue = voxels[ijk[2], ijk[1], ijk[0]]  # note that numpy array index order is kji (not ijk)

Modify voxels in a volume

Typically the fastest and simplest way of modifying voxels is by using numpy operators. Voxels can be retrieved in a numpy array using the array method and modified using standard numpy methods. For example, threshold a volume:

nodeName = "MRHead"
thresholdValue = 100
voxelArray = array(nodeName) # get voxels as numpy array
voxelArray[voxelArray < thresholdValue] = 0 # modify voxel values
getNode(nodeName).Modified() # at the end of all processing, notify Slicer that the image modification is completed

This example shows how to change voxels values of the MRHead sample volume. The values will be computed by function f(r,a,s,) = (r-10)*(r-10)+(a+15)*(a+15)+s*s.

volumeNode=slicer.util.getNode("MRHead")
ijkToRas = vtk.vtkMatrix4x4()
volumeNode.GetIJKToRASMatrix(ijkToRas)
imageData=volumeNode.GetImageData()
extent = imageData.GetExtent()
for k in range(extent[4], extent[5]+1):
  for j in range(extent[2], extent[3]+1):
    for i in range(extent[0], extent[1]+1):
      position_Ijk=[i, j, k, 1]
      position_Ras=ijkToRas.MultiplyPoint(position_Ijk)
      r=position_Ras[0]
      a=position_Ras[1]
      s=position_Ras[2]
      functionValue=(r-10)*(r-10)+(a+15)*(a+15)+s*s
      imageData.SetScalarComponentFromDouble(i,j,k,0,functionValue)
imageData.Modified()

Get volume voxel coordinates from markup fiducial RAS coordinates

This example shows how to get voxel coordinate of a volume corresponding to a markup fiducial point position.

# Inputs
volumeNode = getNode("MRHead")
markupsNode = getNode("F")
markupsIndex = 0

# Get point coordinate in RAS
point_Ras = [0, 0, 0, 1]
markupsNode.GetNthFiducialWorldCoordinates(markupsIndex, point_Ras)

# If volume node is transformed, apply that transform to get volume's RAS coordinates
transformRasToVolumeRas = vtk.vtkGeneralTransform()
slicer.vtkMRMLTransformNode.GetTransformBetweenNodes(None, volumeNode.GetParentTransformNode(), transformRasToVolumeRas)
point_VolumeRas = transformRasToVolumeRas.TransformPoint(point_Ras[0:3])

# Get voxel coordinates from physical coordinates
volumeRasToIjk = vtk.vtkMatrix4x4()
volumeNode.GetRASToIJKMatrix(volumeRasToIjk)
point_Ijk = [0, 0, 0, 1]
volumeRasToIjk.MultiplyPoint(np.append(point_VolumeRas,1.0), point_Ijk)
point_Ijk = [ int(round(c)) for c in point_Ijk[0:3] ]

# Print output
print(point_Ijk)

Get markup fiducial RAS coordinates from volume voxel coordinates

This example shows how to get position of maximum intensity voxel of a volume (determined by numpy, in IJK coordinates) in RAS coordinates so that it can be marked with a markup fiducial.

# Inputs
volumeNode = getNode("MRHead")
markupsNode = getNode("F")

# Get voxel position in IJK coordinate system
import numpy as np
volumeArray = slicer.util.arrayFromVolume(volumeNode)
# Get position of highest voxel value
point_Kji = np.where(volumeArray == volumeArray.max())
point_Ijk = [point_Kji[2][0], point_Kji[1][0], point_Kji[0][0]]

# Get physical coordinates from voxel coordinates
volumeIjkToRas = vtk.vtkMatrix4x4()
volumeNode.GetIJKToRASMatrix(volumeIjkToRas)
point_VolumeRas = [0, 0, 0, 1]
volumeIjkToRas.MultiplyPoint(np.append(point_Ijk,1.0), point_VolumeRas)

# If volume node is transformed, apply that transform to get volume's RAS coordinates
transformVolumeRasToRas = vtk.vtkGeneralTransform()
slicer.vtkMRMLTransformNode.GetTransformBetweenNodes(volumeNode.GetParentTransformNode(), None, transformVolumeRasToRas)
point_Ras = transformVolumeRasToRas.TransformPoint(point_VolumeRas[0:3])

# Add a markup at the computed position and print its coordinates
markupsNode.AddFiducial(point_Ras[0], point_Ras[1], point_Ras[2], "max")
print(point_Ras)

Get the values of all voxels for a label value

If you have a background image called ‘Volume’ and a mask called ‘Volume-label’ created with the Editor you could do something like this:

import numpy
volume = array("Volume")
label = array("Volume-label")
points  = numpy.where( label == 1 )  # or use another label number depending on what you segmented
values  = volume[points] # this will be a list of the label values
values.mean() # should match the mean value of LabelStatistics calculation as a double-check
numpy.savetxt("values.txt", values)

Access values in a DTI tensor volume

This example shows how to access individual tensors at the voxel level.

First load your DWI volume and estimate tensors to produce a DTI volume called ‘Output DTI Volume’

Then open the python window: View->Python interactor

Use this command to access tensors through numpy:

tensors = array("Output DTI Volume")

Type the following code into the Python window to access all tensor components using vtk commands:

volumeNode=slicer.util.getNode("Output DTI Volume")
imageData=volumeNode.GetImageData()
tensors = imageData.GetPointData().GetTensors()
extent = imageData.GetExtent()
idx = 0
for k in range(extent[4], extent[5]+1):
  for j in range(extent[2], extent[3]+1):
    for i in range(extent[0], extent[1]+1):
      tensors.GetTuple9(idx)
      idx += 1

Change window/level (brightness/contrast) or colormap of a volume

This example shows how to change window/level of the MRHead sample volume.

volumeNode = getNode("MRHead")
displayNode = volumeNode.GetDisplayNode()
displayNode.AutoWindowLevelOff()
displayNode.SetWindow(50)
displayNode.SetLevel(100)

Change color mapping from grayscale to rainbow:

displayNode.SetAndObserveColorNodeID("vtkMRMLColorTableNodeRainbow")

Make mouse left-click and drag on the image adjust window/level

In older Slicer versions, by default, left-click and drag in a slice view adjusted window/level of the displayed image. Window/level adjustment is now a new mouse mode that can be activated by clicking on its toolbar button or running this code:

slicer.app.applicationLogic().GetInteractionNode().SetCurrentInteractionMode(slicer.vtkMRMLInteractionNode.AdjustWindowLevel)

Reset field of view to show background volume maximized

Equivalent to click small rectangle button (“Adjust the slice viewer’s field of view…”) in the slice view controller.

slicer.util.resetSliceViews()

Rotate slice views to volume plane

Aligns slice views to volume axes, shows original image acquisition planes in slice views.

volumeNode = slicer.util.getNode("MRHead")
layoutManager = slicer.app.layoutManager()
for sliceViewName in layoutManager.sliceViewNames():
  layoutManager.sliceWidget(sliceViewName).mrmlSliceNode().RotateToVolumePlane(volumeNode)

Iterate over current visible slice views, and set foreground and background images

slicer.util.setSliceViewerLayers(background=mrVolume, foreground=ctVolume)

Internally, this method performs something like this:

layoutManager = slicer.app.layoutManager()
for sliceViewName in layoutManager.sliceViewNames():
   compositeNode = layoutManager.sliceWidget(sliceViewName).sliceLogic().GetSliceCompositeNode()
   # Setup background volume
   compositeNode.SetBackgroundVolumeID(mrVolume.GetID())
   # Setup foreground volume
   compositeNode.SetForegroundVolumeID(ctVolume.GetID())
   # Change opacity
   compositeNode.SetForegroundOpacity(0.3)

Show a volume in slice views

Recommended:

volumeNode = slicer.util.getNode("YourVolumeNode")
slicer.util.setSliceViewerLayers(background=volumeNode)

or

Show volume in all visible views where volume selection propagation is enabled:

volumeNode = slicer.util.getNode("YourVolumeNode")
applicationLogic = slicer.app.applicationLogic()
selectionNode = applicationLogic.GetSelectionNode()
selectionNode.SetSecondaryVolumeID(volumeNode.GetID())
applicationLogic.PropagateForegroundVolumeSelection(0)

or

Show volume in selected views:

n =  slicer.util.getNode("YourVolumeNode")
for color in ["Red", "Yellow", "Green"]:
  slicer.app.layoutManager().sliceWidget(color).sliceLogic().GetSliceCompositeNode().SetForegroundVolumeID(n.GetID())

Change opacity of foreground volume in slice views

slicer.util.setSliceViewerLayers(foregroundOpacity=0.4)

or

Change opacity in a selected view

lm = slicer.app.layoutManager()
sliceLogic = lm.sliceWidget("Red").sliceLogic()
compositeNode = sliceLogic.GetSliceCompositeNode()
compositeNode.SetForegroundOpacity(0.4)

Turning off interpolation

You can turn off interpolation for newly loaded volumes with this script from Steve Pieper.

def NoInterpolate(caller,event):
  for node in slicer.util.getNodes("*").values():
    if node.IsA("vtkMRMLScalarVolumeDisplayNode"):
      node.SetInterpolate(0)

slicer.mrmlScene.AddObserver(slicer.mrmlScene.NodeAddedEvent, NoInterpolate)

You can place this code snippet in your .slicerrc.py file to always disable interpolation by default.

Running an ITK filter in Python using SimpleITK

Open the “Sample Data” module and download “MR Head”, then paste the following snippet in Python interactor:

import SampleData
import SimpleITK as sitk
import sitkUtils

# Get input volume node
inputVolumeNode = SampleData.SampleDataLogic().downloadMRHead()
# Create new volume node for output
outputVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode", "MRHeadFiltered")

# Run processing
inputImage = sitkUtils.PullVolumeFromSlicer(inputVolumeNode)
filter = sitk.SignedMaurerDistanceMapImageFilter()
outputImage = filter.Execute(inputImage)
sitkUtils.PushVolumeToSlicer(outputImage, outputVolumeNode)

# Show processing result
slicer.util.setSliceViewerLayers(background=outputVolumeNode)

More information:

Get axial slice as numpy array

An axis-aligned (axial/sagittal/coronal/) slices of a volume can be extracted using simple numpy array indexing. For example:

import SampleData
volumeNode = SampleData.SampleDataLogic().downloadMRHead()
sliceIndex = 12

voxels = slicer.util.arrayFromVolume(volumeNode)  # Get volume as numpy array
slice = voxels[sliceIndex:,:]  # Get one slice of the volume as numpy array

Get reformatted image from a slice viewer as numpy array

Set up red slice viewer to show thick slab reconstructed from 3 slices:

sliceNodeID = "vtkMRMLSliceNodeRed"

# Get image data from slice view
sliceNode = slicer.mrmlScene.GetNodeByID(sliceNodeID)
appLogic = slicer.app.applicationLogic()
sliceLogic = appLogic.GetSliceLogic(sliceNode)
sliceLayerLogic = sliceLogic.GetBackgroundLayer()
reslice = sliceLayerLogic.GetReslice()
reslicedImage = vtk.vtkImageData()
reslicedImage.DeepCopy(reslice.GetOutput())

# Create new volume node using resliced image
volumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLScalarVolumeNode")
volumeNode.SetIJKToRASMatrix(sliceNode.GetXYToRAS())
volumeNode.SetAndObserveImageData(reslicedImage)
volumeNode.CreateDefaultDisplayNodes()
volumeNode.CreateDefaultStorageNode()

# Get voxels as a numpy array
voxels = slicer.util.arrayFromVolume(volumeNode)
print(voxels.shape)

Combine multiple volumes into one

This example combines two volumes into a new one by subtracting one from the other.

import SampleData
[input1Volume, input2Volume] = SampleData.SampleDataLogic().downloadDentalSurgery()

import slicer.util
a = slicer.util.arrayFromVolume(input1Volume)
b = slicer.util.arrayFromVolume(input2Volume)

# ``a`` and ``b`` are numpy arrays,
# they can be combined using any numpy array operations
# to produce the result array ``c``
c = b - a

volumeNode = slicer.modules.volumes.logic().CloneVolume(input1Volume, "Difference")
slicer.util.updateVolumeFromArray(volumeNode, c)
setSliceViewerLayers(background=volumeNode)

Add noise to image

This example shows how to add simulated noise to a volume.

import SampleData
import numpy as np

# Get a sample input volume node
volumeNode = SampleData.SampleDataLogic().downloadMRHead()

# Get volume as numpy array and add noise
voxels = slicer.util.arrayFromVolume(volumeNode)
voxels[:] = voxels + np.random.normal(0.0, 20.0, size=voxels.shape)
slicer.util.arrayFromVolumeModified(volumeNode)

Mask volume using segmentation

This example shows how to blank out voxels of a volume outside all segments.

# Input nodes
volumeNode = getNode("MRHead")
segmentationNode = getNode("Segmentation")

# Write segmentation to labelmap volume node with a geometry that matches the volume node
labelmapVolumeNode = slicer.mrmlScene.AddNewNodeByClass("vtkMRMLLabelMapVolumeNode")
slicer.modules.segmentations.logic().ExportVisibleSegmentsToLabelmapNode(segmentationNode, labelmapVolumeNode, volumeNode)

# Masking
import numpy as np
voxels = slicer.util.arrayFromVolume(volumeNode)
mask = slicer.util.arrayFromVolume(labelmapVolumeNode)
maskedVoxels = np.copy(voxels)  # we don't want to modify the original volume
maskedVoxels[mask==0] = 0

# Write masked volume to volume node and show it
maskedVolumeNode = slicer.modules.volumes.logic().CloneVolume(volumeNode, "Masked")
slicer.util.updateVolumeFromArray(maskedVolumeNode, maskedVoxels)
slicer.util.setSliceViewerLayers(maskedVolumeNode)

Apply random deformations to image

This example shows how to apply random translation, rotation, and deformations to a volume to simulate variation in patient positioning, soft tissue motion, and random anatomical variations. Control points are placed on a regularly spaced grid and then each control point is displaced by a random amount. Thin-plate spline transform is computed from the original and transformed point list.

https://gist.github.com/lassoan/428af5285da75dc033d32ebff65ba940

Thick slab reconstruction and maximum/minimum intensity volume projections

Set up red slice viewer to show thick slab reconstructed from 3 slices:

sliceNode = slicer.mrmlScene.GetNodeByID("vtkMRMLSliceNodeRed")
appLogic = slicer.app.applicationLogic()
sliceLogic = appLogic.GetSliceLogic(sliceNode)
sliceLayerLogic = sliceLogic.GetBackgroundLayer()
reslice = sliceLayerLogic.GetReslice()
reslice.SetSlabModeToMean()
reslice.SetSlabNumberOfSlices(10) # mean of 10 slices will computed
reslice.SetSlabSliceSpacingFraction(0.3) # spacing between each slice is 0.3 pixel (total 10 * 0.3 = 3 pixel neighborhood)
sliceNode.Modified()

Set up red slice viewer to show maximum intensity projection (MIP):

sliceNode = slicer.mrmlScene.GetNodeByID("vtkMRMLSliceNodeRed")
appLogic = slicer.app.applicationLogic()
sliceLogic = appLogic.GetSliceLogic(sliceNode)
sliceLayerLogic = sliceLogic.GetBackgroundLayer()
reslice = sliceLayerLogic.GetReslice()
reslice.SetSlabModeToMax()
reslice.SetSlabNumberOfSlices(600) # use a large number of slices (600) to cover the entire volume
reslice.SetSlabSliceSpacingFraction(0.5) # spacing between slices are 0.5 pixel (supersampling is useful to reduce interpolation artifacts)
sliceNode.Modified()

The projected image is available in a vtkImageData object by calling reslice.GetOutput().

Display volume using volume rendering

qSlicerAbstractCoreModule* volumeRenderingModule =
  qSlicerCoreApplication::application()->moduleManager()->module("VolumeRendering");
vtkSlicerVolumeRenderingLogic* volumeRenderingLogic =
  volumeRenderingModule ? vtkSlicerVolumeRenderingLogic::SafeDownCast(volumeRenderingModule->logic()) : 0;
vtkMRMLVolumeNode* volumeNode = mrmlScene->GetNodeByID('vtkMRMLScalarVolumeNode1');
if (volumeRenderingLogic)
  {
  vtkSmartPointer<vtkMRMLVolumeRenderingDisplayNode> displayNode =
    vtkSmartPointer<vtkMRMLVolumeRenderingDisplayNode>::Take(volumeRenderingLogic->CreateVolumeRenderingDisplayNode());
  mrmlScene->AddNode(displayNode);
  volumeNode->AddAndObserveDisplayNodeID(displayNode->GetID());
  volumeRenderingLogic->UpdateDisplayNodeFromVolumeNode(displayNode, volumeNode);
  }
logic = slicer.modules.volumerendering.logic()
volumeNode = slicer.mrmlScene.GetNodeByID('vtkMRMLScalarVolumeNode1')
displayNode = logic.CreateVolumeRenderingDisplayNode()
displayNode.UnRegister(logic)
slicer.mrmlScene.AddNode(displayNode)
volumeNode.AddAndObserveDisplayNodeID(displayNode.GetID())
logic.UpdateDisplayNodeFromVolumeNode(displayNode, volumeNode)

Apply a custom volume rendering color/opacity transfer function

vtkColorTransferFunction* colors = ...
vtkPiecewiseFunction* opacities = ...
vtkMRMLVolumeRenderingDisplayNode* displayNode = ...
vtkMRMLVolumePropertyNode* propertyNode = displayNode->GetVolumePropertyNode();
propertyNode->SetColor(colorTransferFunction);
propertyNode->SetScalarOpacity(opacities);
// optionally set the gradients opacities with SetGradientOpacity

Volume rendering logic has utility functions to help you create those transfer functions: SetWindowLevelToVolumeProp, SetThresholdToVolumeProp, SetLabelMapToVolumeProp.

Limit volume rendering to a specific region of the volume

vtkMRMLAnnotationROINode]* roiNode =...
vtkMRMLVolumeRenderingDisplayNode* displayNode = ...
displayNode->SetAndObserveROINodeID(roiNode->GetID());
displayNode->SetCroppingEnabled(1);
displayNode.SetAndObserveROINodeID(roiNode.GetID())
displayNode.CroppingEnabled = True

Register a new Volume Rendering mapper

You need to derive from vtkMRMLVolumeRenderingDisplayNode and register your class within vtkSlicerVolumeRenderingLogic.

void qSlicerMyABCVolumeRenderingModule::setup()
{
  vtkMRMLThreeDViewDisplayableManagerFactory::GetInstance()->
    RegisterDisplayableManager("vtkMRMLMyABCVolumeRenderingDisplayableManager");

  this->Superclass::setup();

  qSlicerAbstractCoreModule* volumeRenderingModule =
    qSlicerCoreApplication::application()->moduleManager()->module("VolumeRendering");
  if (volumeRenderingModule)
    {
    vtkNew<vtkMRMLMyABCVolumeRenderingDisplayNode> displayNode;
    vtkSlicerVolumeRenderingLogic* volumeRenderingLogic =
      vtkSlicerVolumeRenderingLogic::SafeDownCast(volumeRenderingModule->logic());
    volumeRenderingLogic->RegisterRenderingMethod(
      "My ABC Volume Rendering", displayNode->GetClassName());
    }
  else
    {
    qWarning() << "Volume Rendering module is not found";
    }
}

If you want to expose control widgets for your volume rendering method, then register your widget with addRenderingMethodWidget().

Register custom volume rendering presets

Custom presets can be added to the volume rendering module by calling AddPreset() method of the volume rendering module logic. The example below shows how to define multiple custom volume rendering presets in an external MRML scene file and add them to the volume rendering module user interface.

Create a MyPresets.mrml file that describes two custom volume rendering presets:

<MRML version="Slicer4.4.0">
  <VolumeProperty id="vtkMRMLVolumeProperty1" name="MyPreset1"     references="IconVolume:vtkMRMLVectorVolumeNode1;" interpolation="1" shade="1" diffuse="0.66" ambient="0.1" specular="0.62" specularPower="14" scalarOpacity="10 -3.52844023704529 0 56.7852325439453 0 79.2550277709961 0.428571432828903 415.119384765625 1 641 1" gradientOpacity="4 0 1 160.25 1" colorTransfer="16 0 0 0 0 98.7223 0.196078431372549 0.945098039215686 0.956862745098039 412.406 0 0.592157 0.807843 641 1 1 1" />
  <VectorVolume id="vtkMRMLVectorVolumeNode1" references="storage:vtkMRMLVolumeArchetypeStorageNode1;" />
  <VolumeArchetypeStorage id="vtkMRMLVolumeArchetypeStorageNode1" fileName="MyPreset1.png"  fileListMember0="MyPreset1.png" />

  <VolumeProperty id="vtkMRMLVolumeProperty2" name="MyPreset2"     references="IconVolume:vtkMRMLVectorVolumeNode2;" interpolation="1" shade="1" diffuse="0.66" ambient="0.1" specular="0.62" specularPower="14" scalarOpacity="10 -3.52844023704529 0 56.7852325439453 0 79.2550277709961 0.428571432828903 415.119384765625 1 641 1" gradientOpacity="4 0 1 160.25 1" colorTransfer="16 0 0 0 0 98.7223 0.196078431372549 0.945098039215686 0.956862745098039 412.406 0 0.592157 0.807843 641 1 1 1" />
  <VectorVolume id="vtkMRMLVectorVolumeNode2" references="storage:vtkMRMLVolumeArchetypeStorageNode2;" />
  <VolumeArchetypeStorage id="vtkMRMLVolumeArchetypeStorageNode2" fileName="MyPreset2.png"  fileListMember0="MyPreset2.png" />
</MRML>

For this example, thumbnail images for the presets should be located in the same directory as MyPresets.mrml, with the file names MyPreset1.png and MyPreset2.png.

Use the following code to read all the custom presets from MyPresets.mrml and load it into the scene:

presetsScenePath = "MyPresets.mrml"

# Read presets scene
customPresetsScene = slicer.vtkMRMLScene()
vrPropNode = slicer.vtkMRMLVolumePropertyNode()
customPresetsScene.RegisterNodeClass(vrPropNode)
customPresetsScene.SetURL(presetsScenePath)
customPresetsScene.Connect()

# Add presets to volume rendering logic
vrLogic = slicer.modules.volumerendering.logic()
presetsScene = vrLogic.GetPresetsScene()
vrNodes = customPresetsScene.GetNodesByClass("vtkMRMLVolumePropertyNode")
vrNodes.UnRegister(None)
for itemNum in range(vrNodes.GetNumberOfItems()):
  node = vrNodes.GetItemAsObject(itemNum)
  vrLogic.AddPreset(node)