An open-source toolkit, led by Kitware, Inc., for the segmentation, registration, and analysis of tubes and surfaces in images.
C++ CMake Python Shell CSS C XSLT
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.
Applications Merge pull request #949 from aylward/ScaleSkewVersorIn2DAnd3D Jul 25, 2018

TubeTK: Tubular Object Extraction, Registration, and Analysis

Circle CI

TubeTK is an open-source toolkit for the segmentation, registration, and analysis of tubes and surfaces in images, developed by Kitware, Inc.

Tubes and surfaces, as generalized 1D and 2D manifolds in N-dimensional images, are essential components in a variety of image analysis tasks. Instances of tubular structures in images include blood vessels in magnetic resonance angiograms and b-mode ultrasound images, wires in microscopy images of integrated circuits, roads in aerial photographs, and nerves in confocal microscopy.

A guiding premise of TubeTK is that by focusing on 1D and 2D manifolds we can devise methods that are insensitive to the modality, noise, contrast, and scale of the images being analyzed and to the arrangement and deformations of the objects in them. In particular, we propose that TubeTK's manifold methods offer improved performance for many applications, compared to methods involving the analysis of independent geometric measures (e.g., edges and corners) or requiring complete shape models.

TubeTK offers various interface layers:

  • TubeTK/Base: This is the algorithms library. It is the lowest level of access to the methods of TubeTK. It is only available via C++, and it requires considerable expertise to effectively combine and call its methods to do anything useful. Interfacing directly with these algorithms is not recommended and is not well supported. Unit-level testing is performed continuously on these methods.

  • TubeTK/ITKModules: This is the ITK interface to select methods in TubeTK/Base. This level of interface is intended for ITK users and Python scripts writers. The methods exposed represent a level of modularization that invites experimentation, integration with other toolkits (e.g., Scikit-Learn), and development of processing pipelines that accomplish significant image analysis goals. The interface is available as an ITK Extension and thereby available via Python using Wrapped ITK.

  • TubeTK/Applications: These are the command-line interface (CLI) equivalents to the methods available via TubeTK/ITKModules. This is intended for bash, bat, and other system-call scripts. The level of modularization and intended users are similar to those of TubeTK/ITKModules. C++ and python-based CLIs are provided. Continuous, unit-level testing of TubeTK/ITKModules is provided via these applications.

  • TubeTK/Experiments: These are Python Jupyter notebooks that combine many TubeTK/ITKModules into Python scripts that show how to accomplish high-level image analysis goals with TubeTK. They are intended to be an interactive basis for exploring TubeTK. Python and Jupyter notebooks packages must be installed on your computer to run these. These can also be (and are) run as tests to check performance (these test performance, whereas the unit-level tests focus on regression).

Python requirements

To be able to run all the python tests and examples, the following packages are required:

  • numpy
  • scipy
  • jupyter
  • IPython
  • tornado
  • pyzmq
  • jinja2
  • tables
  • matplotlib
  • pyqtgraph
  • PyOpengl
  • PySide
  • ITK

Installing most required packages can be done with the following command line:

pip install requirements.txt

To install ITK, please refer to the ITK documentation or define the environment variables ITK_BUILD_DIR and TubeTK_BUILD_DIR to point respectivally to the folder in which ITK and TubeTK are built.


TubeTK is licensed under the Apache License, Version 2.0 (the "License"); you may not use TubeTK except in compliance with the License. Unless required by applicable law or agreed to in writing, TubeTK is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.


The development of TubeTK is supported in part by the