# Transform Module

**URL:** https://discourse.slicer.org/t/transform-module/18667
**Category:** Support
**Created:** [July 13, 2021, 3:17pm UTC](https://discourse.slicer.org/t/transform-module/18667 "2021-07-13T15:17:49Z")
**Posts on this page:** 4
**Page:** 1

<div class="post-metadata">

### Author: ![reddington.15](https://avatars.discourse-cdn.com/v4/letter/r/34f0e0/32.png) [@reddington.15](https://discourse.slicer.org/u/reddington.15)
#### Post date: [July 13, 2021, 3:17pm UTC](https://discourse.slicer.org/t/transform-module/18667/1 "2021-07-13T15:17:49Z")

</div>

Hello,

In a brief synopsis I am trying to calculate the deformation of the pelvis from a pre and post CT scan after impact testing. Once I segment both pelvis bones out using segmentation editor and wrap solidfy, I align the segments using fiducial registration by placing points on each segment and then transform. I also then use segmentation registration, which gives me more transformation matrices. I just want to make sure I am understanding the matrices I am obtaining correctly. When I align with fiducial registration and then transform is that matrix the path of the points on the post to align with the pre? Then segmentation registration is the deformation between the segments once overlapped and that deformation is calculated in the deformable matrix or the affine matrix? I hope this makes sense and that I have given you enough information, but if not I am happy to add clarification.

Thanks!

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [July 13, 2021, 6:37pm UTC](https://discourse.slicer.org/t/transform-module/18667/2 "2021-07-13T18:37:03Z")

</div>

It is not necessary to use both the point based registration and segment registration - one should be enough.

> [@reddington.15](#):
>
> When I align with fiducial registration and then transform is that matrix the path of the points on the post to align with the pre?

The transform moves the “From” (or “Moving”) object to the “To” (or “Fixed”) object.

> [@reddington.15](#):
>
> Then segmentation registration is the deformation between the segments once overlapped and that deformation is calculated in the deformable matrix or the affine matrix?

You can choose to compute rigid, affine, or deformable matrix. The deformable transform is always applied on top of a linear bulk transform (otherwise the matrix inverse computation would much less stable).

---

<div class="post-metadata">

### Author: ![reddington.15](https://avatars.discourse-cdn.com/v4/letter/r/34f0e0/32.png) [@reddington.15](https://discourse.slicer.org/u/reddington.15)
#### Post date: [July 19, 2021, 3:24am UTC](https://discourse.slicer.org/t/transform-module/18667/3 "2021-07-19T03:24:11Z")

</div>

Okay, that sort of makes sense. Is there a reason why the compute rigid, affine, or deformable matrix are slightly different in value?

---

<div class="post-metadata">

### Author: ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)
#### Post date: [July 19, 2021, 11:44am UTC](https://discourse.slicer.org/t/transform-module/18667/4 "2021-07-19T11:44:41Z")

</div>

The transforms are different in what displacements they can describe.

- Rigid = translation + rotation.
- Affine = translation + rotation + shear.
- Deformable: any smoothly changing displacement field.

For bones, you usually rigid transform (if you register multiple bones then piecewise rigid) is the most suitable, unless they are warped.
