MIRTK 2.0

After a long period of continuous development, the next major release of MIRTK is now available!

This release version was used in the experiments of the PhD thesis of Andreas Schuh, and the "Unbiased construction of a temporally consistent morphological atlas of neonatal brain development", preprint available on bioRxiv (doi: https://doi.org/10.1101/251512).

Please note that the reason for an increase of major version number is not only the substantial added functionality since the previous release, but mainly because the scaling of the image dissimilarity gradient of the Registration module has been changed. In consequence, the weights of regularization terms have to be adjusted to account for this change, i.e., the registration will give different results than with previous versions.

Alternatively, consider adding the following options to your mirtk register configuration file:

• Normalize image gradient = False (cf. 992ccdb)
• Precompute image derivatives = False if Interpolation mode used is either Linear [with padding] or the default Fast linear [with padding] mode (cf. a68aefb).

MIRTK for dHCP pipeline 1.0

This MIRTK release is used in the structural-pipeline of the @DevelopingHCP.

MIRTK 1.1

We are pleased to announce the release of MIRTK 1.1.

This release gets rid of bugs, runs on Linux, OS X, and Windows, and provides a Python package for execution of MIRTK commands in a pipeline script. The commands now support the NIfTI-2 image file format for large images and the GIFTI file format for cortical surface meshes as used by the HCP project. The build system has further improved with CMake BASIS 3.3 and automated builds help to keep the project stable.

An Ubuntu 14.04 Docker image with MIRTK pre-installed is available from Docker Hub; the quickest way to getting to know MIRTK or to perform reproducible experiments in a confined environment!

See ChangeLog for a complete list of changes.

Start using MIRTK 1.1 today!

MIRTK 1.0

First stable release of the Medical Image Registration ToolKit (MIRTK).

The MIRTK is the official successor of the Image Registration Toolkit (IRTK) developed at the Biomedical Image Analysis (BioMedIA) group at Imperial College London. The state of the art image and point set registration methods of the IRTK have been completely rewritten in a modular and extensible framework, which mirrors the formulation of the registration problem as energy minimisation with analytic gradient calculations.

Additional packages included in the MIRTK can be used to reconstruct cortical surface meshes given a tissue segmentation and to inflate those surfaces. The volumetric mapping package is used to find a volumetric map from one domain to another, such as from a given brain volume to the unit ball. These tools are in active development and will be extended in the coming months.

See ChangeLog and online documentation for more.

MIRTK 1.0rc2

This is the second release candidate of the Medical Image Registration ToolKit (MIRTK):

• Fixed initialization of object factories when linked statically or dynamically with GNU g++.
• Build NIfTI C Library and libLBFGS as part of the modules that depend on them instead.
• Added separate Image I/O module providing all image reader/writer classes.

See ChangeLog and online documentation for more.

MIRTK 1.0rc1

This is the first public release candidate of the Medical Image Registration ToolKit (MIRTK).

See ChangeLog and online documentation for more.