SCI Seg3D Mailing List

Text archives Help

Re: [Seg3D] Smoothing Isosurfaces after Segmentation

Chronological Thread 
  • From: Jess Tate <>
  • Cc:
  • Subject: Re: [Seg3D] Smoothing Isosurfaces after Segmentation
  • Date: Mon, 11 Apr 2016 11:59:23 -0600

Hi Daniel,

Paul's suggestion is a good one, as MeshLab does have more mesh manipulation tools than SCIRun.  I would like to add that Seg3D will export isosurfaces as stl files, which MeshLab should be able to read.  

SCIRun can also probably do what you need.  To get your isosurface into SCIRun, the ascii or stl format is better.  Even better still is to save out your segmentation as a nrrd and load that into SCIRun and make an Isosurface there (GetDomainBoundary or Extract(Simple)Isosurface).  You can use fairmesh to smooth your Isosurfaces, and as long as they do not overlap in your segmentation (they will need at least full pixel in between boundaries, even diagonally.), then the isosurfaces will not overlap.  However, fairmesh does not guarantee any sort of quality of elements, so the result may not be useful for computation.  It can help you make better meshes with distance maps though.  

You can consider using some meshing tools we have too.  Cleaver will make multimaterial volume meshes, but you can use getDomainboundary to get the boundary surfaces.  

Another tip is that most meshing tools that go from segmentation to mesh can fail easily with thin structures, but using higher resolutions (use the resample tool in seg3d) can help, though it will increase the computation cost significantly.  Or by you can try meshing the regions independently, which you may have tried already.  

Let me know if you need more info on any of these steps.


On Apr 10, 2016, at 6:07 PM, Paul Holcomb <> wrote:


We use MeshLab to do our isosurface postprocessing, and not only does it work very well, it's free.  I would suggest using the boolean filters in Seg3D to ensure no initial overlap between your surfaces, and then export to vtk.  Unfortunately, MeshLab doesn't support VTK, but one of my colleagues wrote a simple vtk2obj converter in Python, which I've included.  Let me know if you have any questions.

On Sun, Apr 10, 2016 at 8:00 PM DANIEL PFEIFFER-KELLY <> wrote:

---------------- Forwarded Message ----------------
Date: Sun, Apr 10, 2016 03:16 PM
Subject: Smoothing Isosurfaces after Segmentation

To whom it may concern,
Hello my name is Dan Pfeiffer-Kelly a junior mechanical engineering student at Penn State University. I am currently conducting biomechanics and imaging research with one of my professor's (Daniel Cortes). After reading a blog regarding smoothing isosurfaces after segmentation (,  I wanted to ask a few questions. Any help would be greatly appreciated! 
The primary goal of our research is to develop a multi-scale model of the lower leg. I wont go into all the details, but I have attached to this email a more in depth goals and objectives page if you would like to read it over for any clarifications. Right now I am attempting to segment and generate an isosurface for each muscle in the lower leg. However each muscle must be smooth with no overlap between each isosurface. The primary problem I have come across is I have neither been unable to properly smooth the muscle isosurfaces nor know how to prevent the smoothed surfaces from overlapping. I have tried a few different methods including built in smoothing tools in Seg3D (smooth binary dilate/erode) and exporting an STL file to SketchUp and editing it there. However, none of the methods have produced the results I am looking for. I was wondering if you had any advice? What would be the most practical method to smoothing isosurfaces while still allowing the two surfaces to fit together without overlap or significant gaps? 
After reading over the blog I mentioned above,  the software SCIRun came to my attention. However, I have been having trouble uploading the isosurfaces to SCIRun and using the program in general. I was wondering if you could please clarify how to upload the files to SCIRun and smooth them there? I am unsure if it is just my macbook causing the problem. I plan to try the software on my colleagues PC. I hope to hear back soon and thanks so much for any incite.
Daniel Pfeiffer-Kelly


Seg3d mailing list:
To unsubscribe, email with the "unsubscribe seg3d" in the message body.
Seg3d mailing list:
To unsubscribe, email with the "unsubscribe seg3d" in the message body.

Archive powered by MHonArc 2.6.18.

Top of page