# control rigidity in bspline registration

**URL:** https://discourse.itk.org/t/control-rigidity-in-bspline-registration/4380
**Category:** Uncategorized
**Tags:** registration, features, bspline
**Created:** [August 29, 2021, 4:46am UTC](https://discourse.itk.org/t/control-rigidity-in-bspline-registration/4380 "2021-08-29T04:46:41Z")
**Posts on this page:** 2
**Page:** 1

<div class="post-metadata">

### Author: ![mary\_na](https://discourse.itk.org/letter_avatar_proxy/v4/letter/m/7c8e57/32.png) [@mary\_na](https://discourse.itk.org/u/mary_na)
#### Post date: [August 29, 2021, 4:46am UTC](https://discourse.itk.org/t/control-rigidity-in-bspline-registration/4380/1 "2021-08-29T04:46:41Z")

</div>

Hello! In general the cost function of registration optimization could have two part. One for similarity measurement and the other to control rigidity (registration penalty). In itk I just can change similarity metric. Is there any feature in itk to change registration penalty?

Thanks in advance

---

<div class="post-metadata">

### Author: ![zivy](https://discourse.itk.org/user_avatar/discourse.itk.org/zivy/32/1726_2.png) [@zivy](https://discourse.itk.org/u/zivy)
#### Post date: [August 29, 2021, 6:47pm UTC](https://discourse.itk.org/t/control-rigidity-in-bspline-registration/4380/2 "2021-08-29T18:47:21Z")

</div>

Hello @mary_na,

What you are referring to is a regularization term (not specific to rigid), which is not available. The similarity metrics implemented in the ITK registration framework only use the image similarity and do not include a term for transformation parameter regularization.

That being said, the standalone Demons family of algorithms does include regularization via smoothing (elastic or viscous).