Computational Morphometry Toolkit (CMTK) News

CMTK 2.3.0 released

Torsten Rohlfing — Fri, 17 May 2013 1:28:23 GMT

We are pleased to announce the release of CMTK 2.3.0, which brings a number of bug fixes as well as a vastly improved tool for detection and quantification of the ADNI imaging phantom. This will work even in many cases when the phantom itself is defective.

CMTK distributions now also include the SRI24 atlas for use with some bundled image analysis pipelines.

See Release Notes and ChangeLog for full details.

CMTK 2.2 release adds new features

Torsten Rohlfing — Tue, 08 May 2012 11:28:19 GMT

We are pleased to announce the release of CMTK 2.2 with some exciting new features. Source code and pre-compiled packages are available from CMTK's "Downloads" section on NITRC.

Most importantly, CMTK now has two tools to perform structural MRI unwarping using reference scans of the "ADNI Phantom." The workflow is described in a brief paper over at the InsightJournal:

http://hdl.handle.net/10380/3354

This article can also be downloaded from the CMTK/Docs section on NITRC.

Many other detail fixes and improvements have also gone into this release - check out the ChangeLog for details.

As always, all of CMTK is free as in beer and speech, and you don't even have to make us authors on any of your papers just for using it. Imagine that! (Please do mention CMTK in your Acknowledgments if you use it, though.)

[Note: the original 2.2.0 release had a packaging bug and was replaced with release 2.2.1]

CMTK 2.1.3 minor release

Torsten Rohlfing — Fri, 09 Mar 2012 12:57:41 GMT

We have released CMTK 2.1.3. This release is intended to lock in minor changes and reorganization of the Shape-Based Averaging (SBA) tools and library code, formerly contained in the "average_edt" tool.

Users of CMTK 2.1.2 who are not interested specifically in SBA probably have no reason to upgrade.

Of interest for users on the Mac platform may be the addition of pre-built packages compiled using the MacPorts gcc 4.6 compiler. These builds, indicated by an additional "mp" in the package file name, replace SMP parallelism using GCD with OpenMP parallelism. On dedicated computation servers, this should give substantially improved performance of some CMTK tools, because only a small fraction of algorithms is implemented using GCD support.

CMTK 2.1.2 bugfix release

Torsten Rohlfing — Thu, 23 Feb 2012 6:12:14 GMT

We have released CMTK 2.1.2, which fixes a compilation problem using the latest version of zlib (1.2.6). Also fixes build problems when using MacPorts on MacOS, where MacPorts zlib is not compatible with system zlib.

New functionality in this release includes library classes and command line tools for fitting low-dimensional to high-dimensional transformations (e.g., affine to B-spline or B-spline to deformation field).

CMTK 2.1.1 released

Torsten Rohlfing — Wed, 18 Jan 2012 3:58:52 GMT

We are pleased to announce the release of CMTK 2.1.1.

Most significant addition in this release is a new tool, "epiunwarp", for correction of B0-field distortion in echo-planar MR images (e.g., diffusion-weighted images). The tool implements an algorithm by Holland et al. (NeuroImage 50(1):175-183, 2010; http://dx.doi.org/10.1016/j.neuroimage.2009.11.044) that requires two EPI acquisitions with opposing phase encoding directions, but no field map.

This release also adds a simple tool ("gmm") for tissue classification using a Gaussian mixture model and atlas-based tissue priors.

CMTK now in NeuroDebian

Torsten Rohlfing — Mon, 12 Dec 2011 12:35:36 GMT

The CMTK Team is proud to announce CMTK's availability through NeuroDebian, starting with Release 2.1.0 of the toolkit.

If you are already using NeuroDebian, simply "update" and select cmtk for installation, e.g., from the command line, do the following:

sudo apt-get update
sudo apt-get install cmtk

Users who do not have the NeuroDebian repository linked yet will need to go to

http://neuro.debian.net/#how-to-use-this-repository

and select their release and desired mirror to use (note that cmtk may not have propagated to all mirrors yet). Afterwards, to install CMTK from the command line run

sudo apt-get install cmtk

CMTK 1.7.0 Released

Torsten Rohlfing — Mon, 09 May 2011 8:13:31 GMT

We are pleased to announce the release of version 1.7.0 of CMTK, the Computational Morphometry Toolkit.

This release greatly improves the functionality of CMTK on the MacOS and Windows (native) platforms. For shared-memory parallelism, OpenMP is now fully supported using MS Visual C++, and Grand Central Dispatch is supported on MacOS.

See Release Notes and ChangeLog for full details.

CMTK is made available under the GPLv3 with support from NIBIB (Grant No. EB008381)

CMTK 1.4.0 adds support for GPU computing using CUDA

Torsten Rohlfing — Tue, 13 Jul 2010 3:36:59 GMT

We are pleased to announce Release 1.4.0 of CMTK, The Computational Morphometry Toolkit. This release adds support for GPU computation using CUDA, currently only for MR intensity bias field correction. Additional tools will receive GPU support in upcoming versions.

CMTK 1.1.0 released with Slicer support

Torsten Rohlfing — Thu, 27 Aug 2009 2:17:35 GMT

Version 1.1.0 of CMTK, the Computational Morphometry Toolkit, is now available.

This release brings the first tools that fully support the command line plugin interface of 3D Slicer. Among these tools are affine and B-spline nonrigid image registration, correction of interleaved image motion artifacts, and MR bias field correction.

A new command line tools, also available as Slicer plugin, now performs integrated atlas-based image segmentation.

Finally, we fixed a major bug in the native Windows build, which effectively broke multi-threaded computation on that platform.

For more details on changes in this release, please see the change log.

CMTK 1.0.0 released

Torsten Rohlfing — Wed, 17 Jun 2009 11:13:21 GMT

We are pleased to announce CMTK 1.0.0, the first full open source release of the Computational Morphometry Toolkit.

This is the first complete release of the entire toolkit under the GPL, which will be the basis for all future development of the software.

Pre-compiled packages are provided for Linux (Fedora 11; 32bit), Cygwin, MacOS (PPC) and Windows. Additional binary packages for different Linux flavors and MacOS on Intel are being built and will be added as they arrive.