Rendering meshes in VTK

Deborah Schmidt

Helmholtz Imaging | MDC Berlin

Sep 25, 2024

Slides available at https://ida-mdc.github.io/workshop-visualization/tutorial-mesh-rendering-vtk/

Introduction

Python based 3D visualization tool
Supports various formats (STL, OBJ, PLY, ..)
Advanced rendering capabilities (volume rendering, surface rendering, light and material configuration)
Supports programmatic animations

Schroeder, Will; Martin, Ken; Lorensen, Bill (2006), The Visualization Toolkit (4th ed.), Kitware, ISBN 978-1-930934-19-1, https://vtk.org/

Solution to run VTK code

To simplify and unify solutions used in our tutorials, we use Album, a tool for capturing and sharing specific software use cases in dedicated virtual environments.

Click the solution box next to this text and follow the displayed usage instructions to run the solution either from command line or graphical interface.

Visualize meshes in VTK

visualization:visualize-meshes-vtk:0.1.0

Displays meshes and optionally renders them to video.

Close solution

Visualize meshes in VTK

visualization:visualize-meshes-vtk:0.1.0

Displays meshes and optionally renders them to video.

License: MIT
Solution creator(s): Deborah Schmidt

Usage:

album install visualization:visualize-meshes-vtk:0.1.0
album run visualization:visualize-meshes-vtk:0.1.0
    --meshes # (directory): Path to the meshes. [Default: ] [Required: Yes]
    --volume # (file): Optional volume to render next to the meshes. [Default: ] [Required: No]
    --include # (string): List of names of elements which should be loaded, comma separated [Default: ] [Required: No]
    --exclude # (string): List of names of elements which should not be loaded, comma separated [Default: ] [Required: No]
    --colormap # (string): Name of the matplotlib colormap to use for coloring meshes. [Default: viridis] [Required: No]
    --output_mp4 # (file): If provided, the result is not visualized in a window, but rendered to this path. [Default: ] [Required: No]
    --size # (string): Size of the video in WxH (i.e. "800x600") [Default: 1600x900] [Required: No]
    --exclude_zero # (boolean): Exclude 0 as a label when processing the volumes. [Default: 1] [Required: No]

Cite the following resources if you use this solution:

Schroeder, Will; Martin, Ken; Lorensen, Bill (2006), The Visualization Toolkit (4th ed.), Kitware, ISBN 978-1-930934-19-1

Dependencies

Link to full source code

316    dependencies={'environment_file': """channels:
317  - conda-forge
318  - defaults
319dependencies:
320  - python=3.9
321  - pandas=1.5.3
322  - matplotlib-base=3.7.0
323  - numpy=1.24.2
324  - vtk=9.2.6
325  - numpy-stl=3.0.1
326  - opencv=4.8.0
327  - pip
328  - pip:
329    - vtkplotlib==2.1.0
330    - imagecodecs
331    - tifffile==2023.9.26
332"""}

Setting the scene

The renderer: Manages the 3D scene by adding objects (called actors), processing lighting, and handling the camera to control how the scene is displayed.
Central component in VTK: The renderer is responsible for compositing all elements (meshes, volumes, etc.) and rendering them into a final image or interactive display.

Link to full source code

97    renderer = vtk.vtkRenderer()

Loading and Adding Meshes

Load STL files via vtkSTLReader
Coloring Meshes via actor.GetProperty().SetColor()

Link to full source code

119        reader = vtk.vtkSTLReader()
120        reader.SetFileName(vtk_file)
121        reader.Update()
122        # Create mapper and actor
123        mapper = vtk.vtkPolyDataMapper()
124        mapper.SetInputConnection(reader.GetOutputPort())
125
126        actor = vtk.vtkActor()
127        actor.SetMapper(mapper)
128
129        color_rgba = cmap(norm(i))
130        actor.GetProperty().SetColor(color_rgba[:3])
131
132        renderer.AddActor(actor)

Loading and Adding Voxel Volumes

Numpy to VTK Conversion

Link to full source code

45    from tifffile import imread
46    import numpy as np
47    import vtk
48    from vtk.util import numpy_support
49
50    tiff_data = imread(volume).astype(np.uint16)
51    # Calculating min and max values from tiff_data
52    min_val = tiff_data.min()
53    max_val = tiff_data.max()
54
55    imageData = vtk.vtkImageData()
56    imageData.SetDimensions(tiff_data.shape)
57    imageData.SetSpacing((1, 1, 1))
58
59    # Efficiently converting the whole numpy array to VTK array
60    vtk_array = numpy_support.numpy_to_vtk(num_array=tiff_data.ravel(), deep=True, array_type=vtk.VTK_UNSIGNED_CHAR)
61    imageData.GetPointData().SetScalars(vtk_array)

Loading and Adding Voxel Volumes

Color and Opacity Transfer Functions

Link to full source code

63    # Setting up color and opacity transfer functions
64    colorTransferFunction = vtk.vtkColorTransferFunction()
65    colorTransferFunction.AddRGBPoint(min_val, 0, 0, 0)  # Mapping min_val to black
66    colorTransferFunction.AddRGBPoint(max_val, 1, 1, 1)  # Mapping max_val to white
67
68    opacityTransferFunction = vtk.vtkPiecewiseFunction()
69    opacityTransferFunction.AddPoint(min_val, 0.0)  # Fully transparent at min_val
70    opacityTransferFunction.AddPoint(max_val, 0.5)  # Semi-transparent at max_val
71
72    volumeProperty = vtk.vtkVolumeProperty()
73    volumeProperty.SetColor(colorTransferFunction)
74    volumeProperty.SetScalarOpacity(opacityTransferFunction)

Loading and Adding Voxel Volumes

Creating a Volume Node

Link to full source code

76    volume = vtk.vtkVolume()
77    volumeMapper = vtk.vtkSmartVolumeMapper()
78    volumeMapper.SetInputData(imageData)
79    volume.SetProperty(volumeProperty)
80    volume.SetMapper(volumeMapper)

Creating the Render Window

Render Window: Manages the actual window where the scene is displayed.
RenderWindowInteractor: Allows for interactive control of the 3D scene using mouse and keyboard inputs.

Link to full source code

33    renderWindow = vtk.vtkRenderWindow()
34    renderWindow.AddRenderer(renderer)
35
36    if output_mp4:
37        render_video(output_mp4, rendering_size, renderWindow, renderer)
38    else:
39        render_window_interactor = vtk.vtkRenderWindowInteractor()
40        render_window_interactor.SetRenderWindow(renderWindow)
41        render_window_interactor.Start()

Result