# Mandible registration

**URL:** <https://discourse.slicer.org/t/mandible-registration/9902>\
**Category:** SlicerCMF\
**Created:** [January 22, 2020, 2:25am UTC](https://discourse.slicer.org/t/mandible-registration/9902 "2020-01-22T02:25:25Z")\
**Posts on this page:** 4\
**Page:** 2

<div class="post-metadata">

**Author:** ![Lexus\_H](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lexus_h/32/4110_2.png) [@Lexus\_H](https://discourse.slicer.org/u/Lexus_H)\
**Post date:** [January 31, 2020, 12:07am UTC](https://discourse.slicer.org/t/mandible-registration/9902/21 "2020-01-31T00:07:23Z")

</div>

![Screen Shot 2020-01-30 at 6.45.07 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/d/a/dac8f82ec179ad0a1e1e27a807af8dcce8aa3ed4.jpeg)

This is surface registration which I could not get it to register well. Please notice the transparent model that did not superimpose well.

 ![Screen Shot 2020-01-30 at 6.50.31 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/a/8/a8692615a16740d2332eaec2320c0f1f870f0a83.jpeg)

This is a registered post defect model with CMF fiducial registration which I believe, works better than ROI. Do I have to see transparent gray for pre defect model?

Dr. P, yes I manually approximated before I segmented both images then created STL for surface registration, in this case, I used fiducial registration.

Even Prof. Lassoan stated that I should use surface registration but somehow surface registration is not giving me a good result. Also, I even clipped to make the two models more even at the borders(distal to molars and mesial to anteriors)

Hi Manjula, I will try general registration using DICOM files. your recommended video clip does not have voice recorded so it is kind of hard for me to follow. Also, should I use slicer CMF for general registration or just slicer?

Can you also explain to me what cloud compare is?

Many thanks,

If anyone is willing to try with dicom files to evaluate whether my work step was OK or not, please let me know, I can transfer them via google drive link. Since I sent my stls and my registered and clipped model and model to model is simply plugging your models in order to generate vtk file for SPV, I do not know what I’m doing wrong. Only the periodontal defect is the change and that should be colored based on the depth and allocated color in the color bar not the rest of the mandible (teeth can move bc it is dry mandible but not as much as defect sizes up to 9mm).

---

<div class="post-metadata">

**Author:** ![Lexus\_H](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lexus_h/32/4110_2.png) [@Lexus\_H](https://discourse.slicer.org/u/Lexus_H)\
**Post date:** [January 31, 2020, 12:13am UTC](https://discourse.slicer.org/t/mandible-registration/9902/22 "2020-01-31T00:13:38Z")

</div>

Hi Manjula,

Are you able to send me the surface registered model to me so I can play around with it to see if the periodontal bone defect matches my gold standard?

Since I just realized that your maximum was -4 to 4, it colored the same as red.

I believe you have my email address from the google drive link.

Thanks,

Erin

 ![Screen Shot 2020-01-30 at 6.45.07 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/d/a/dac8f82ec179ad0a1e1e27a807af8dcce8aa3ed4.jpeg)

 ![Screen Shot 2020-01-30 at 6.50.31 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/a/8/a8692615a16740d2332eaec2320c0f1f870f0a83.jpeg)

---

<div class="post-metadata">

**Author:** ![Lexus\_H](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lexus_h/32/4110_2.png) [@Lexus\_H](https://discourse.slicer.org/u/Lexus_H)\
**Post date:** [January 31, 2020, 3:24pm UTC](https://discourse.slicer.org/t/mandible-registration/9902/23 "2020-01-31T15:24:14Z")

</div>

Hi Manjula,

This is a color map page using slicer CMF 5.0 that i generated using Manjula’s registered models using cloud compare. Not much better as you can see.

 ![color map but no defects shown.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/2/5/2531ed8880e91ddff8f4c41289e406ca78c1e514.jpeg)

 ![Screen Shot 2020-01-30 at 6.45.07 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/d/a/dac8f82ec179ad0a1e1e27a807af8dcce8aa3ed4.jpeg)

 ![Screen Shot 2020-01-30 at 6.50.31 PM.png](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/a/8/a8692615a16740d2332eaec2320c0f1f870f0a83.jpeg)

---

<div class="post-metadata">

**Author:** ![manjula](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/manjula/32/80981_2.png) [@manjula](https://discourse.slicer.org/u/manjula)\
**Post date:** [January 31, 2020, 3:33pm UTC](https://discourse.slicer.org/t/mandible-registration/9902/24 "2020-01-31T15:33:10Z")

</div>

Well i think you are not doing it properly somewhere because this is what you should see if done correctly

Also you need to change the attribute to signed and also the figure you see is just a scale not in millimeters or anything.  
x 0.5 is the center (Superimposed) and to either side you get the deviation by the percentage.

if the range is for e.g -5 to 5 in mm then at ever 0.1 position you will move 1 mm + or -

[Previous page](https://discourse.slicer.org/t/mandible-registration/9902.md?page=1)