Poly Rigid Registration

Summary

Registers a moving image to a fixed image using vertebra-level associations and poly-rigid deformation. Use this for pairwise alignment of two CT (or compatible) images at the spine level; the algorithm can compute or reuse SpineData and optimize alignment along spine splines.

See also: SpineData.

Compatible data

  • Requires: SharedImageSet fixed and moving; optionally SpineData for each.

  • Implicit assumptions: consistent geometry/coordinate systems between images; vertebra associations are meaningful for the chosen mode.

Output

  • Registered moving image; updated SpineData for fixed/moving as needed.

  • Deformation state accessible via the controller/UI.

Key parameters

  • Association Mode (enum): Automatic / Classification / SplineBasedRigid / PolyRigid.

  • Current Offset (0..1): Relative offset along splines for alignment.

  • Num Samples (int): Samples for similarity optimization.

Workflow

  1. Provide fixed and moving images (optionally spines).

  2. Choose an association mode; automatic works in most cases.

  3. Generate SpineData if absent.

  4. Optimize similarity; inspect the plot and select the best offset.

Controller / UI

  • The registration controller provides a plot view for similarity across offsets and exposes alignment parameters.

Tips

  • If association seems wrong, switch mode or correct spines before optimizing.

  • Ensure both images use consistent geometry for best results.