# Segmentation smooth

**URL:** <https://discourse.slicer.org/t/segmentation-smooth/26051>\
**Category:** Support\
**Created:** [November 3, 2022, 12:37pm UTC](https://discourse.slicer.org/t/segmentation-smooth/26051 "2022-11-03T12:37:00Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![qod\_rec](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/qod_rec/32/15572_2.png) [@qod\_rec](https://discourse.slicer.org/u/qod_rec)\
**Post date:** [November 3, 2022, 12:37pm UTC](https://discourse.slicer.org/t/segmentation-smooth/26051/1 "2022-11-03T12:37:00Z")

</div>

Hello! I have a dataset of chest CTs. Here is what I see when I open them with 3d slicer:

 ![ask1.PNG](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/5/f/5f7259b85322707606ee57e196f3d1b9bdf72f23.jpeg)

This segmentation has some sawtoothes both in the label map and the mesh.  
I want to get smooth contour and I saw some videos use the ‘Gaussian smooth’ in ‘Segment Editor’.  
Is this a nice way to maintain the segmentations’ information? Since they are made by radiologists.  
Also the ‘Gaussian smooth’ has the ‘standard deviation’ option. How should I set this in a reasonable way? How many milimeter should I choose?

![ask2](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/e/7/e79291925d8f82f8fe062d9f54cf941aa4d96c53.png)

Thank you!

---

<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:** [December 1, 2022, 8:00am UTC](https://discourse.slicer.org/t/segmentation-smooth/26051/2 "2022-12-01T08:00:53Z")

</div>

The segmentation’s resolution looks very coarse. I don’t think you can get a very accurate segmentation out of this, but if you [increase the resolution of the segmentation](https://slicer.readthedocs.io/en/latest/user_guide/modules/segmenteditor.html#segmentation-is-not-accurate-enough) before applying smoothing then you might get better results.

However, I’m not sure if this is worth it, as you can get much higher quality segmentations using [LungCTAnalyzer extension](https://github.com/rbumm/SlicerLungCTAnalyzer) or [TotalSegmentator extension](https://github.com/lassoan/SlicerTotalSegmentator) - fully automatically, in a few minutes, including splitting the lung to lobes (and also segmenting 100 other structures).

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/2/e/2e49b51e9ed2f5075d5083cef185983aa5ee6a47.jpeg)