SimpleITK now has registration!
The community is encouraged to use and test this registration interface and provide feedback on the design and features.
This beta features the ImageRegistrationMethod which brings a SimpleITK style interface to the modular ITKv4 registration framework. This adds support for a variety of transforms including affine, b-spline, and deformation fields. The metrics supported include correlation, means squares, ANTS neighborhood correlation, and mutual information. A variety of optimizers are available along with scales estimators for the optimized transformation parameters and built in multi-scale registration support.
Additionally, a number of registration filters have been added:
Several examples can be found in the examples directory to help get started. These examples include Affine registration, BSpline, Demons and Displacement fields.
The following filters were also added:
Note: We ran into a limitation with Visual Studio that prevents successful building and packaging for Windows 64. This will be addressed before the final release. Please try the 32-bit Windows version for smaller images.
Before the final release, the goal is to improve testing, add more examples and documentation, and fix bugs. The beta is feature complete for registration work, and reports of missing critical features are welcome.
To install this beta release, manually download the binaries, then explicitly install the downloaded beta file. Binary distributions for many platforms and languages are available for download. See also the nightly Doxygen documentation and additional release notes.