inspire | Spatial Information-Based Deformable Image Registration

inspire is a C++ library. inspire has no bugs, it has no vulnerabilities, it has a Permissive License and it has low support. You can download it from GitHub.

INSPIRE is a deformable image registration method (and tool) for 2D and 3D images. @article{ofverstedt2020inspire, title={INSPIRE: Intensity and Spatial Information-Based Deformable Image Registration}, author={{"O}fverstedt, Johan and Lindblad, Joakim and Sladoje, Nata{\v{s}}a}, journal={arXiv preprint arXiv:2012.07208}, year={2020} }. Please cite this paper if you use INSPIRE (and please check back for an updated peer-reviewed reference in the future). The implementation is based on ITK, a general-purpose library for image processing and analysis, optimisation, registration, segmentation and much more. To build INSPIRE, follow these steps: Download and install ITK (github.com/InsightSoftwareConsortium/ITK). Clone the inspire repo, into, e.g. my_path/inspire Create a new directory my_path/inspire-build From my_path/inspire-build do: "cmake -DCMAKE_BUILD_TYPE=Release ../inspire/inspire_itk" make. Now you are ready to register images (assuming that they have already been registered up to affine transformations). From inspire-build for 2d images, call INSPIRE: ./InspireRegister 2 -ref my_reference_image.tif -flo my_floating_image.tif -deform_cfg ../inspire/inspire_itk/config/default_configuration.json \ -out_path_deform_forward tforward.txt -out_path_deform_reverse treverse.txt. For 3d images: ./InspireRegister 3 -ref my_reference_image.tif -flo my_floating_image.tif -deform_cfg ../inspire/inspire_itk/config/default_configuration.json \ -out_path_deform_forward tforward.txt -out_path_deform_reverse treverse.txt. To apply the transformation to images, use the following command: ./InspireTransform -dim 2 -16bit 1 -interpolation linear -transform tforward.txt -ref my_reference_image.tif -in my_floating_image.tif -out my_output_image.tif. Supported interpolation modes: (nearest, linear, cubic). INSPIRE is configured by editing configuration files written in the standard JSON format. It has an hierarchical structure. For each level of the configuration file, you can specify some preprocessing settings that should be applied to the image, and general settings. Then in each sublevel of the configuration you can specify the number of control-points in the B-spline grid, iterations, learning rate, momentum. Here is a default configuration which goes up to 48 control points in the final level of the registration process, starting out at only 9.