# Demonstration of the point-to-plane metric for registration

**URL:** <https://discourse.vtk.org/t/demonstration-of-the-point-to-plane-metric-for-registration/12517>\
**Category:** Development\
**Created:** [October 20, 2023, 5:43pm UTC](https://discourse.vtk.org/t/demonstration-of-the-point-to-plane-metric-for-registration/12517 "2023-10-20T17:43:26Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Jens\_Munk\_Hansen](https://discourse.vtk.org/user_avatar/discourse.vtk.org/jens_munk_hansen/32/1215_2.png) [@Jens\_Munk\_Hansen](https://discourse.vtk.org/u/Jens_Munk_Hansen)\
**Post date:** [October 20, 2023, 5:43pm UTC](https://discourse.vtk.org/t/demonstration-of-the-point-to-plane-metric-for-registration/12517/1 "2023-10-20T17:43:26Z")

</div>

Hi Developers

I have made a small demo project (overnight) demonstrating registration of `vtkPolyData` using the point-to-plane metric. I have made this following the `VTK` directory structure.

The metric requires either normals on the target or they are estimated similarly to how the `vtkImplicitPolyDataDistance` works. It is the most basic version of ICP point-to-plane, but the intention is that more sophisticated methods will follow, e.g. using the Gauss-Newton scheme, a bi-directional metric, priors and re-weighted least-squares.

This is a few classes that I have extracted from a much larger repository that I am working on.

The code is pretty well-documented, but it is quite far from a new remote module for `VTK`.

> **[GitHub - JensMunkHansen/IterativeClosestPoint: VTK Remote module for...](https://github.com/JensMunkHansen/IterativeClosestPoint)**
>
> VTK Remote module for registration using iterative closest point - GitHub - JensMunkHansen/IterativeClosestPoint: VTK Remote module for registration using iterative closest point

What are your thoughts  
Jens Munk Hansen