# How to use ITK to segment bones in osteoporotic patients

**URL:** <https://discourse.itk.org/t/how-to-use-itk-to-segment-bones-in-osteoporotic-patients/6181>\
**Category:** Algorithms\
**Tags:** segmentation, cpp, registration\
**Created:** [September 6, 2023, 2:10pm UTC](https://discourse.itk.org/t/how-to-use-itk-to-segment-bones-in-osteoporotic-patients/6181 "2023-09-06T14:10:02Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![AeBoPi](https://discourse.itk.org/letter_avatar_proxy/v4/letter/a/6bbea6/32.png) [@AeBoPi](https://discourse.itk.org/u/AeBoPi)\
**Post date:** [September 6, 2023, 2:10pm UTC](https://discourse.itk.org/t/how-to-use-itk-to-segment-bones-in-osteoporotic-patients/6181/1 "2023-09-06T14:10:02Z")

</div>

Osteoporosis reduces bone density and thus the strength of bone in the CT data set (shown in the figure), making it impossible to use traditional thresholding techniques to handle segmentation.  
 ![dd7ddd16096dc1f7b237bb300ebe866](https://discourse.itk.org/uploads/default/original/2X/a/a0bb7ca35e60cafbae1124abb13e6d89ccce4b07.png)

 ![df05a03a46ae758e9f3021b793355a4](https://discourse.itk.org/uploads/default/original/2X/0/011994953cfe5bd3c76297138c72e2469c81de4c.png)

I noticed that the Morphon registration algorithm mentioned in [this paper](https://ieeexplore.ieee.org/abstract/document/1699682)  
seems to work for this case. There is a corresponding [ITK implementation](https://www.insight-journal.org/browse/publication/320), but the [source code](https://github.com/midas-journal/midas-journal-320/tree/master/Morphon) it provides has a version mismatch with the current ITK.

I wonder if there is any other possible way in ITK to deal with this situation? Is DisplacementFieldTransform a viable approach? Displacement fields are also mentioned in Morphon, but I didn’t find an example of applying DisplacementFieldTransform, so I turn to you for help.

---

<div class="post-metadata">

**Author:** ![dzenanz](https://discourse.itk.org/user_avatar/discourse.itk.org/dzenanz/32/1093_2.png) [@dzenanz](https://discourse.itk.org/u/dzenanz)\
**Post date:** [September 6, 2023, 3:11pm UTC](https://discourse.itk.org/t/how-to-use-itk-to-segment-bones-in-osteoporotic-patients/6181/2 "2023-09-06T15:11:26Z")

</div>

You could use [elastix](https://elastix.lumc.nl/), possibly via its [ITKElastix](https://github.com/InsightSoftwareConsortium/ITKElastix) interface.

---

<div class="post-metadata">

**Author:** ![AeBoPi](https://discourse.itk.org/letter_avatar_proxy/v4/letter/a/6bbea6/32.png) [@AeBoPi](https://discourse.itk.org/u/AeBoPi)\
**Post date:** [September 6, 2023, 4:12pm UTC](https://discourse.itk.org/t/how-to-use-itk-to-segment-bones-in-osteoporotic-patients/6181/3 "2023-09-06T16:12:28Z")

</div>

Thank you! I will try.