# Auto3dgm landmark placement on invisible parts of mesh 

**URL:** <https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803>\
**Category:** Support\
**Created:** [January 16, 2026, 10:11pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803 "2026-01-16T22:11:47Z")\
**Posts on this page:** 14\
**Page:** 1

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 16, 2026, 10:11pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/1 "2026-01-16T22:11:47Z")

</div>

Hello Slicer Community,

I’m using Auto3dgm on segmented upper beaks from MarkMyBird database, but it appears landmarks are being placed on seemingly invisible parts of the segmented meshes!

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/f/c/fce004b0ad5ff2acab82120c1f00088f2a7a74dd.jpeg)

Here are the original unsegmented meshes:

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/1/7/17d4d18c855db5fb97880e90b71a8d616b90ca1f.jpeg)

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/3/b/3b7c15a9353d2d0a9609a7982d91ecff58f2a16c.jpeg)

Here’s how I segmented the meshes: I used Plane Cut tool in Dynamic Modeler, using a 2D plane placed vertically at approximately where the beak begins. I exported the desired segment as an .obj file. Interestingly, using either DeCAL or BUILDER module as part of ATLAS, they seem to respect the posterior boundary of the mesh (but they’re not without other issues with these meshes):

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/b/7/b7a456c78d47136d7ab62b911ef20990d64c798d.jpeg)

Any help would be greatly appreciated!

Best,

Ricardo Ely

---

<div class="post-metadata">

**Author:** ![muratmaga](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/muratmaga/32/3622_2.png) [@muratmaga](https://discourse.slicer.org/u/muratmaga)\
**Post date:** [January 16, 2026, 11:10pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/2 "2026-01-16T23:10:08Z")

</div>

I think that extension is abandoned, I don’t think anyone in the development team tracks this forum. You might try opening an issue on their github repo [GitHub - ToothAndClaw/SlicerAuto3dgm: Repo for auto3dgm slicer extension](https://github.com/ToothAndClaw/SlicerAuto3dgm) but they don’t appear to be responsive.

I think if you use Dynamic modeler to cut a model, it removes the vertices but not the points. We encountered. that in other application. I opened an issue, [Dynamic modeler ROI cut removes the polygons but not the vertices · Issue #8777 · Slicer/Slicer · GitHub](https://github.com/Slicer/Slicer/issues/8777)

@lassoan maybe it is time to implement that clean option.

---

<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:** [January 17, 2026, 1:27pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/3 "2026-01-17T13:27:16Z")

</div>

OK, I’ll add the clean option in the next few days.

---

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 17, 2026, 6:05pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/4 "2026-01-17T18:05:58Z")

</div>

@lassoan That would be amazing, 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:** [January 18, 2026, 3:10pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/5 "2026-01-18T15:10:56Z")

</div>

I’ve pushed a fix, it is under review.

> <https://github.com/Slicer/SlicerSurfaceToolbox/pull/85>
>
> Dynamic Modeler module's ROI Cut tool and Plane Cut tool use vtkClipPolyData, wh…ich can leave orphan points (points that are not used in any cells) in its output. The points are not exactly the same as the input points (coordinates are, but point IDs are not) and they are only preserved in certain cases (e.g., when clipped points are requested), therefore the orphan points are not reliable usable for anything and so they are not worth preserving. This is confirmed by testing the filter and also discussed at https://discourse.vtk.org/t/vtkclippolydata-has-different-behaviors-in-vtk-8-2-0-and-in-vtk-9-4-0/15626.
> 
> Since having orphan points in the output is unexpected and caused errors in subsequent processing, clean polydata filter is applied to the output to remove the orphan points.
> 
> fixes https://github.com/Slicer/Slicer/issues/8777

---

<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:** [January 19, 2026, 2:16am UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/6 "2026-01-19T02:16:29Z")

</div>

The fix is integrated. Slicer Preview Releases (downloaded tomorrow or later) will not have invisible orphan points in the outputs of ROI cut and Plane cut tools.

---

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 20, 2026, 10:59am UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/7 "2026-01-20T10:59:15Z")

</div>

Thank you for implementing those fixes! I was just about to download the preview release for Windows, but it seems that version is from 01/15, while Linux and mac are 01/19? I run 3D Slicer on Windows.

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/d/4/d4890d13b6a7d98323d515a341ace2f6d86bab8d.png)

---

<div class="post-metadata">

**Author:** ![cpinter](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/cpinter/32/7995_2.png) [@cpinter](https://discourse.slicer.org/u/cpinter)\
**Post date:** [January 20, 2026, 12:54pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/8 "2026-01-20T12:54:22Z")

</div>

It seems the last build succeeded on all platforms. Click the package button in the Windows line and you can download the installer (apparently the download site hasn’t updated itself yet).

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/a/5/a5834d500ee6cb9ea07791c0ec6e5cd0b2d4a719.png)

> **[SlicerPreview](https://slicer.cdash.org/index.php?project=SlicerPreview&date=2026-01-20)**
>
> CDash is an open source, web-based software testing server. CDash aggregates, analyzes, and displays the
> results of software testing processes submitted from clients located around the world. CDash is a part of a
> ...

Please keep in mind that for such a recent preview you may need to wait a bit for all the extensions be available.

---

<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:** [January 20, 2026, 6:29pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/9 "2026-01-20T18:29:44Z")

</div>

The download page now shows 2026-01-20 for Slicer Preview Releases for all platforms. This version contains the fix.

---

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 21, 2026, 3:55pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/10 "2026-01-21T15:55:38Z")

</div>

Hello all,

Auto3dgm works now with the new fix. Thank you @lassoan for taking the time to address this issue!

Best,

Ricardo Ely

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/f/4/f41f4934448bb66b743ec08215dbd82f6f6602ec.jpeg)

---

<div class="post-metadata">

**Author:** ![muratmaga](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/muratmaga/32/3622_2.png) [@muratmaga](https://discourse.slicer.org/u/muratmaga)\
**Post date:** [January 21, 2026, 4:21pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/11 "2026-01-21T16:21:13Z")

</div>

If you look at the clipped surface, I think you will see many point there? You might want to consider there having points in a completely arbitrary and flat surface makes biological sense in your geometric analysis.

---

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 21, 2026, 4:33pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/12 "2026-01-21T16:33:31Z")

</div>

I made sure to uncheck the option for capping a surface under the Plane Cut tool, forcing landmarks to fall on the surface mesh.

 ![image](https://us1.discourse-cdn.com/flex002/uploads/slicer/original/3X/3/8/38cb24427c4159be05091c28bc5e527364c96120.jpeg)

---

<div class="post-metadata">

**Author:** ![raranda22](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/raranda22/32/80343_2.png) [@raranda22](https://discourse.slicer.org/u/raranda22)\
**Post date:** [January 22, 2026, 10:02am UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/13 "2026-01-22T10:02:31Z")

</div>

On second thought, regarding how much biological sense my procedure makes, while I made sure to leave the clipped surface uncapped, I’m now slightly concerned about landmarks falling on the clipped edges. I may be overthinking this too much, but the criteria for cutting the mesh along that specific plane is the part of the beak that is anterior to the nasals. The angle of the plane is determined by a PCA of the specimen’s landmarks themselves (for the most part). Any thoughts?

---

<div class="post-metadata">

**Author:** ![muratmaga](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/muratmaga/32/3622_2.png) [@muratmaga](https://discourse.slicer.org/u/muratmaga)\
**Post date:** [January 22, 2026, 4:39pm UTC](https://discourse.slicer.org/t/auto3dgm-landmark-placement-on-invisible-parts-of-mesh/45803/14 "2026-01-22T16:39:06Z")

</div>

I don’t think you will be concerned on points falling on the clipped surface. The question is now consistent and repeatable the clipping is.