VTK
The Visualization ToolKit (VTK) is an open source, freely available software system for 3D computer graphics, image processing, and visualization used by thousands of researchers and developers around the world. VTK consists of a C++ class library, and several interpreted interface layers including Tcl/Tk, Java, and Python. Professional support and products for VTK are provided by Kitware, Inc. (www.kitware.com) VTK supports a wide variety of visualization algorithms including scalar, vector, tensor, texture, and volumetric methods; and advanced modeling techniques such as implicit modelling, polygon reduction, mesh smoothing, cutting, contouring, and Delaunay triangulation. In addition, dozens of imaging algorithms have been directly integrated to allow the user to mix 2D imaging / 3D graphics algorithms and data.
Documentation Improvement
(If you agree with the following, vote for the improvements!!)
In my experience, the single most important factor that influences a users attitude about a software package is its ease of use. VTK provides a phenomenal collection of tools for scientific data processing and visualization. However, if a user has difficulty accessing these tools, their experience with, and therefore opinion of, the software will be less than satisfactory. Documentation is the key element which controls this difficulty level. I propose some improvements to the documentation system here.
Example Usage (C++)
These are fully independent, compilable examples. There is significant overlap in the examples, but they are each intended to illustrate a different concept and be fully stand alone compilable.
Please add examples in your areas of expertise!
System Configuration/General Information
- Which Libraries Do I Link To?
- Which Header Files Do I Include?
- Typical CMakeLists.txt file
- Environment Setup
- [A nice tutorial]
Preliminaries
Wrapping
Simple Operations
- Distance between two points
- Random number (uniform distribution)
- Random number (Gaussian distribution)
Input and Output
Input
- Read a simple "xyz" file of points
- Read a plain text file into a polydata
- Read a delimited file into a polydata
- Read a VTU file
- Read an OBJ File
- Convert a series of DICOM files into a VTI File
- VRML (WRL)
Output
Geometric Objects
In this section, the object is created, added to a polydata object, then written to a .vtp file. The file can be opened in Paraview to see the result.
- Point
- Line
- Add colored lines to a Polydata
- Plane
- Triangle (on large point array)
- Triangle (self contained)
- Sphere
- Cube
- Polygon
Working with 3D Data (PolyData)
Input/Output (Reading/Writing)
Geometry and Topology - Adding Points and Shapes to Polydata
A big confusion among VTK beginners is "I added points to my file - why are there no points when I visualize it??". VTK strongly divides GEOMETRY from TOPOLOGY. What most users would think of as "points" are actually "points + vertices" in VTK. The geometry is ALWAYS simply a list of coordinates - the topology represents the connectedness of these coordinates. If no topology at all is specified, then, as far as VTK is aware, there is NOTHING to show. If you want to see points, you must tell VTK that each point is independent of the rest by creating a vertex (a 0-D topology) at each point. The following examples use a triangle to demonstrate a step at a time how to add geometry and topology to polydata.
- Triangle - Geometry only - Create a polydata consisting of the three corners of a triangle. There is nothing to visualize as there is no topology.
- Triangle - Geometry + Vertices - Create a polydata consisting of the three corners of a triangle. A vertex is added at each point so there is now 3 "points" for the user to see. This is 0-D topology.
- Triangle - Geometry + Lines - Create a polydata consisting of three corners of a triangle. A line is added between each point. This is 1-D topology.
- Triangle - Geometry + Polygon - Create a polydata consisting of three corners of a triangle. A polygon (in this case, a triangle) is added on the three points. This is 2-D topology.
Adding Colors to Polydata
- Colored Points - Add three points to a polydata and associate a color with each of them.
- Triangle - Colored Points - Set the color of each point of a triangle. You will be able to interpolate the colors across the triangle.
- Triangle - Solid Color - Create a solid colored triangle.
Adding Normals to Polydata
- Add normal vectors at each vertexAdd Normals to a Polydata
- Add normal vectors at each point and each vertex (what is the difference?)Point and Cell Normals
Non-standard Data
These examples show how to attach your own, nonstandard fields to every point or cell in a polydata.
- Add Miscellaneous Data to Points in a Polydata
- Get Miscellaneous Data from Points in a Polydata
- Add Miscellaneous Data to Cells in a Polydata
- Get Miscellaneous Data from Cells in a Polydata
Point cloud operations
Miscellaneous (Field) Data
Three types of data can be stored in a polydata object, PointData, CellData, and FieldData. For PointData, there must be a piece of data associated with each point (e.g. a temperature data array with the temperature at each point). For CellData, there must be a piece of data associated with each cell (e.g. the area of each triangle). For data that does not align with either points or cells, FieldData should be used. This is typically data that describes the dataset as a whole. An example could be the name of the dataset, or the center of mass of the points, etc.
- Using FieldData Add Global Miscellaneous Data to a Polydata
Global Operations on PolyData
- Extract Normals from a Polydata
- Get the names of all of the data arrays
- Determine data types of arrays
Utilities
- Color Lookup Table
- Construct a Table - A table is a 2D array of any type of elements. They do not all have to be the same type. This is achieved using vtkVariant.
- Delaunay Triangulation (2D) - Perform a 2D Delaunay triangulation on a point set.
- Compute eigenvalues and eigenvectors of a symmetric matrix
- 2D Array
Graphs
Data Structures
Filters
- Apply a Transformation to Points
- Landmark Transform
- Iterative Closest Points (ICP) Transform
- Triangulate a Terrain Map
- Create a surface from Unorganized Points
- Create a surface from Unorganized Points (Gaussian Splat)
- Create models from labeled volume data (Discrete MarchingCubes)
- Create cubes from labeled volume data
Visualization
See [this] for a brief explanation of the VTK terminology of mappers, actors, etc.
- Visualize a Sphere
- Visualize a VTP File
- Visualize a 2D Set of Points
- Display coordinate axes
- "Trackball" mode
- Display text
Working with Images
- Display a static image - this will display the image, but not allow you to interact with it.
- Visualize and interact with an image
- Select a region of the window
Working with Meshes
- Decimation - Reduce the number of triangles in a mesh.
- Subdivision - Increase the number of triangles in a mesh.
Needed/Missing Examples!
- vtkVolumeRayCastMapper - any example of raycast/volume rendering
Example Usage (Python)
Getting Started
Working with PolyData
- Write the corners of a triangle to a file (python)
- Write the corners of a triangle to a file (+vertices) (python)
- Write a triangle with colored points (python)
- Write a colored triangle (python)
- Write a solid colored triangle (python)
- Write a triangle to a file (python)
- Iterative Closest Points (ICP) (python)
Dashboard submissions
Running tests locally
To run all of the tests on your modified source tree, simply type <source lang="text"> ctest </source> in your build directory. This will not submit anything to the dashboard, but it is a good "first test" to simply see, locally, if everything works.
Submitting an experimental build to the dashboard
The idea of this type of submission is simply to have a nice way to view the output of all the tests, and to leave "proof" that you indeed tested the code. This is useful if you then commit your code and it breaks someone else's build - you can then claim you did everything you could :). To run this type of submission, simply type <source lang="text"> make Experimental </source> from your build directory.
Creating a 'Nightly' dashboard submission
It is impossible for developers to test code on every operating system, compiler, and configuration. By creating a dashboard submission, you can help them find bugs that could be affecting many users but are transparent to some developers. The idea is to get the latest source code, compile it, and run a battery of tests - reporting any compile, build, and test errors to a system which very neatly arranges the results (http://www.cdash.org/CDash/index.php?project=VTK).
It is recommended to not use the same build you work with daily for you dashboard submission. If there is a problem with the nightly cvs, your code may not compile the next day!
To get started, create a new directory called /home/username/Dashboards/VTK. It does not actually have to be in this exactly directory, but this path will be used throughout this example to make the ideas concrete. cd to your new directory and run these commands to check out an initial version of VTK and data sets used for testing.
Here is an example cmake dashboard file.
You will probably want to submit a dashboard every night, so you can add a cronjob. Run 'crontab -e' and enter the following command <source lang="text"> 0 1 * * * export DISPLAY=:0.0 ; ctest -S /home/username/Dashboards/VTK/dashboard.cmake -V > /home/username/Dashboards/VTK/dashboard.log 2>&1 </source>
This says "at 1:00 AM, every day, every month, run the dashboard tests and log verbose output to dashboard.log". The DISPLAY variable is set to allow the tests that need an X server to complete successfully.
Developers Corner
- If you work with Paraview and VTK, the recommended procedure is to use the same source tree for both projects. That is, build Paraview from .../src/Paraview and VTK from .../src/Paraview/VTK.
Pitfalls
- When using the Set*Macro's, you must initialize the variable that is being set in the constructor. The Set*Macro does a check to see if the value has changed before changing it (so that it can update the Modified variable), so if the value is uninitialized, this will cause an error in some memory checking tools (such as valgrind).
Examples for Developers
Wiki Sandbox
Administrative Topics
- Where can I find more information about VTK?
- Where can I download VTK?
- Where can I download a tarball of the nightly HTML documentation?
- Where can I get VTK Datasets?
- VTK Patch Procedure -- merge requests for the current release branch
- What are some projects using VTK?
Current Projects
- VTK Graph Layout
- VTK Java Wrapping
- Composite Data Redesign
- VTK Widget Redesign
- Shaders in VTK
- VTK with Matlab
- VTK Time support
- VTK Depth Peeling
- VTK Multi-Pass Rendering
- Using VTK with JRuby
- Painters
- Cray XT3 Compilation
- Statistics
- Array Refactoring
- Multicore and Streaming
- 3DConnexion Devices Support
External Links
- IMTEK Mathematica Supplement (IMS), the Open Source IMTEK Mathematica Supplement (IMS) interfaces VTK and generates ParaView batch scripts
- [1], VTK examples in C# (Visual Studio 5.0 and .NET 2.0)
Development Process
The VTK Community is upgrading its development process. We are doing this in response to the continuing and rapid growth of the toolkit. A VTK ARB is being put in place to provide strategic guidance to the community, and individuals are being identified as leaders in various VTK subsystems.
