Working with VTK

The Visualization Toolkit (VTK) is an open-source, freely available software system for 3D computer graphics, image processing, and visualization. It consists of a C++ class library and several interpreted interface layers including Tcl/Tk, Java, and Python. VTK supports a wide variety of visualization algorithms including scalar, vector, tensor, texture, and volumetric methods, as well as advanced modeling techniques such as implicit modeling, polygon reduction, mesh smoothing, cutting, contouring, and Delaunay triangulation. VTK has an extensive information visualization framework and a suite of 3D interaction widgets. The toolkit supports parallel processing and integrates with various databases on GUI toolkits such as Qt and Tk. VTK is cross-platform and runs on Linux, Windows, Mac, and Unix platforms. VTK is part of Kitware’s collection of commercially supported open-source platforms for software development.

Installing

VTK version 8.0.0 and later is available on PyPi for python versions 2.7, 3.4, 3.5 and 3.6

$ pip install vtk

On Windows only python 3.5 and 3.6 are supported. VTK can also be installed with anaconda. Available versions are at https://anaconda.org/conda-forge/vtk/files.

If you want to build VTK yourself, download it from https://www.vtk.org. When building toggle the flag VTK_WRAP_PYTHON to generate the wrapping files. Detailed instructions can bee seen here

Wrappers

Reading and writing VTK image files to and from numpy array requires a lot of boilerplate code. The classes below, VTKImage and VTKInformation wraps reading and writing vti files, ie images of type vtkImageData. To read an image, write: image = VTKImage.read('path.vti'). If image will be modified you have to save its attributes: info = image.information().

VTKImage is a subclass of numpy.ndarray and can be used as any normal numpy array.

Finally the image can be written to disk

VTKImage.write_vti('new_path.vti', image, info)
class VTKImage

Numpy ndarray wrapper for VTK images. This class holds meta data about the image such that reading and writing to file will keep the correct attributes

>>> import numpy as np
>>> import dictlearn as dl
>>> # volume is a numpy array
>>> volume = VTKImage.read('path/to/volume.vti')
>>> assert isinstance(volume, np.ndarray)
>>> prod = np.dot(volume[:10, :10], np.random.rand(10, 5))
>>> assert prod.shape == (10, 5)
>>> patches = dl.Patches(volume)
>>> patch_generator = patches.create_batch_and_reconstruct(10000):
>>> for batch, reconstruct in patch_generator:
>>>     # Handle batch
>>>     reconstruct(batch)

Write ‘patches.reconstructed’ to disk with the same attributes as ‘path/to/volume.vti’

>>> volume.write('path/to/volume_new.vti', patches.volume)
information(self)

Get image meta data. See VTKInformation

static from_array(array, info=None)

Create a VTKImage from a numpy array

static from_image_data(image_data, name=None)

Crate VTKImage from vtkImageData

Parameters:
  • image_data – vtk.vtkImageData instance
  • name – Name of point array to extract. Defaults to array at index 0
Returns:

VTKImage
static read(path, name=None)

Read a vti image.

Parameters:
  • path – Path to file
  • name – Name or index of array. If ‘name’ is None then array at index 0 is returned
Returns:

VTKImage instance
write(self, path, array=None)

Write data (array or self) to ‘vti’ file. This file is written with self.extent, self.origin, self.spacing and self.dtype. If the instance is created with VTKImage.read() these attributes are copied from the read file, otherwise the default values are used:

  • extent = [0, self.dimensions[0] - 1,
    0, self.dimensions[1] - 1, 0, self.dimensions[2] - 1]
  • origin = [0, 0, 0]
  • spacing = [1, 1, 1]
  • dtype = np.float64
Parameters:
  • path – Filename, where to save
  • array – Optional, if array is None ‘self’ is written to file. If array is not None then array is written to file
Returns:

True if writing successful
static write_vti(path, array, info=None, extent=None, origin=None, spacing=None, use_array_type=True, name='ImageScalars')

Write ‘array’ to ‘path’ as vti file

Parameters:
  • path – Where to write
  • array – Data to write, ndarray with array.ndim == 3
  • info – Optional instance of VTKInformation, overwrites extent, origin and spacing.
  • extent – Data extent, array like, len(extent) == 6. Default [0, array.shape[0] - 1,0, array.shape[1] - 1, 0, array.shape[2] - 1]
  • origin – Data origin, default [0, 0, 0]
  • spacing – Spacing between voxels, default [1, 1, 1]
  • use_array_type – Only used if info is not None. If this is False the image is saved with the data type given by info, otherwise array.dtype is used
  • name – Name of scalar array
Returns:

True if write successful
print(self)

Print image information

copy(self, order='C')

Return a copy of the image

Parameters:order – {‘C’, ‘F’, ‘A’, ‘K’}, optional Controls the memory layout of the copy. ‘C’ means C-order, ‘F’ means F-order, ‘A’ means ‘F’ if a is Fortran contiguous, ‘C’ otherwise. ‘K’ means match the layout of a as closely as possible.
class VTKInformation(path=None, reader=None)

Holds image metadata

  • datatype, VTK datatype, int
  • bounds, bounds of the geometry, size 6
  • center, center of the geometry, size 3
  • dimensions, size of the geometry, size 6
  • extent, six integers - give the index of the first and last
    point in each direction
  • origin,
  • spacing,
Parameters:
  • path – Path to vtk image
  • reader – Instance of a vtk image reader
vtp_to_vti(surface, information, invalue=1, outvalue=0, flip=None)

Convert a closed surface to ImageData using vtkPolyDataToImageStencil. All points on or inside the takes ‘invalue’ while all point outside the surface takes ‘outvalue’

Parameters:
  • surface – Path to surface file
  • information – Information about the volume to create. Either an instance of VTKInformation or path to a vti file. If this is a path to an image, its attributes are copied to the converted image
  • invalue – Value of points inside or of the surface
  • outvalue – Value of points outside the surface.
  • flip – Flip around an axis, options: ‘x’, ‘y’, ‘z’ or None to keep as is
Returns:

An instance of VTKImage