Skip to content

dMRI based fiber tracking using parallel transport tractography

License

Notifications You must be signed in to change notification settings

LucasBetioli123/trekker

 
 

Repository files navigation

Trekker

Trekker implements a state-of-the-art tractography algorithm, parallel transport tractography (PTT). PTT is capable of reconstructing geometrically smooth and topographically organized fiber bundles.

Trekker software has the following features:

  • Clean probabilistic tractography with PTT.
  • Pathway rules that offer flexible options to reconstruct intricate connections.
  • Time limiting enables tracking for a given length of duration, specification of streamline count is optional.
  • Multithreading reconstructs streamlines using multiple cores of a CPU.
  • Support for asymmetric FODs provides a more flexible option for fiber tracking.
  • Command line interface, an essential requirement to write scripts for processing big data in computer grids.
  • Python package, wrapped using Cython from Trekker's C/C++ code which provides easy access to PTT algorithm for python lovers.
  • .vtk output, that is compatible with a large number of third party 3D rendering software, for easy and high quality visualizations.

For complete documentation, tutorials and examples, visit https://dmritrekker.github.io/.

Example:

#Clean whole brain tractography with intuitive options

./trekker -fod FOD.nii.gz \
          -seed_image WHITEMATTER.nii.gz \
          -seed_count 100000 \
          -pathway=discard_if_ends_inside WHITEMATTER.nii.gz \
          -pathway=discard_if_enters CSF.nii.gz \
          -minLength 20 \
          -maxLength 300 \
          -output OUTPUT.vtk

Installation

Stand-alone executables:

Latest binaries are under: https://github.com/baranaydogan/trekker/tree/master/binaries

Building from source:

Step 1. Download the source code:

git clone https://github.com/dmritrekker/trekker

Step 2. Modify the first few lines in the build script (make sure to have cmake-3.15 or newer):

  • For Linux -> build_Linux.sh
  • For Windows -> build_Windows.bat

Step 3. Run the build script. This will build Trekker under:

  • For Linux -> /build/Linux/install
  • For Windows -> /build/Windows/install

Python package:

A python package can be built from source. You can also download a built distributions (.whl) from the /binaries folder.

To install the python package use:

pip install Trekker-<VERSION>.whl

Demo

To test your installation, please use the scripts under -> /demo

Publications:

If you use Trekker, please cite [Aydogan2021], which provides a detailed explanation of the algorithm and the results obtained from extensive experiments. A precursory abstract was initially published during ISMRM 2019, Montreal [Aydogan2019], which can additionally be cited:

  • [Aydogan2021] Aydogan D.B., Shi Y., "Parallel transport tractography", in IEEE Transactions on Medical Imaging, vol. 40, no. 2, pp. 635-647, Feb. 2021, doi: 10.1109/TMI.2020.3034038.

  • [Aydogan2019] Aydogan D.B., Shi Y., "A novel fiber tracking algorithm using parallel transport frames", Proceedings of the 27th Annual Meeting of the International Society of Magnetic Resonance in Medicine (ISMRM) 2019

About

dMRI based fiber tracking using parallel transport tractography

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • C++ 80.9%
  • C 15.3%
  • Cython 1.6%
  • CMake 1.0%
  • Python 0.7%
  • MATLAB 0.3%
  • Other 0.2%