SCIRun User Mailing List

Text archives Help


Re: [SCIRUN-USERS] Questions about BioMesh3D


Chronological Thread 
  • From: Nima Maftoon <n_maftoon@yahoo.com>
  • To: Jess Tate <jess@sci.utah.edu>
  • Cc: "scirun-users@sci.utah.edu" <scirun-users@sci.utah.edu>
  • Subject: Re: [SCIRUN-USERS] Questions about BioMesh3D
  • Date: Wed, 10 Jun 2015 15:16:54 +0000 (UTC)
Dear Jess and Brett,

Thanks for your great explanations. Let me go back to the itemized format again.

1) I don't seem to find the relevant document pages about mesh conversion. Could you please point me there?

2 & 3)
(a) "You will need to transform it back into the original coordinate system (can be done is scirun). You can also use SCIRun to extract the surfaces from the tet mesh." How? Which module does that? Can a part of a surface be selected to make an element group for applying boundary conditions?
(b) By "transformed "tight" files" do you mean that they are transformed to the correct coordinate system? Are "tight" files the surface mesh files?
(c) "there is a problem doing mix type meshes.  BioMesh/SCIRun does not support mixed mesh type, so they would have to be all surfaces or all volumes.  Additionally,  BioMesh (and almost all meshing software) has problems with thin features." I understand this. FE meshers are perfectly happy with surface meshes and mixing them with volume meshes. This mixing job can be done there. A very thin structure is usually represented by one open line in segmentation because as you implied capturing the thickness is usually very hard if not impossible. Triangulation of these open lines across slices provides the 3D model of that structure. An example of such a structure is the tympanic membrane. In such a structure the segmentation essentially lacks volume information.
4) Does specifying the max volume size, have any effect on the surface element and edge sizes as well?

Thanks
Nima

From: Jess Tate <jess@sci.utah.edu>
To:
Cc: Nima Maftoon <n_maftoon@yahoo.com>; "scirun-users@sci.utah.edu" <scirun-users@sci.utah.edu>
Sent: Tuesday, June 9, 2015 4:40 PM
Subject: Re: [SCIRUN-USERS] Questions about BioMesh3D

Hi Nima,

To add what Brett provided, generating surfaces should be fine, but there is a problem doing mix type meshes.  BioMesh/SCIRun does not support mixed mesh type, so they would have to be all surfaces or all volumes.  Additionally,  BioMesh (and almost all meshing software) has problems with thin features.  Your sampling resolution for your input segmentation will need to be high enough to accurately represent the boundaries, which takes more time.  In terms of thin structures,  If Biomesh can generate a surface mesh, it can usually generate a volume mesh.

cheers,
Jess





On Jun 9, 2015, at 2:29 PM, Brett Burton <zaracay@gmail.com> wrote:

Nima,

Surfaces are generated, but BioMesh transforms the images into a different coordinate space during its initial stage.  It is not transformed back until the final stage.  So, pulling files like particle-union.ts.fld or the tissue1.tight.fld, which are surface meshes, will ultimately be in the wrong space, but (like Jess said) they can be transformed back in SCIRun.  

I have also altered the original python scripts to give the transformed "tight" files if you would like to have those files automatically generated.  I have not created the python scripts for the particle-union files, though it wouldn't take much time to do so if you need them.  I can check those changes in, or send you the scripts that you can use to replace your old python scripts.

Brett

On Jun 9, 2015, at 12:00 PM, Nima Maftoon <n_maftoon@yahoo.com> wrote:

Thanks Jess!

About Question #3: shell elements are essentially surface elements. I am wondering why this should be difficult? I thought in one of the intermediate stages surface elements are generated.

Nima

From: Jess Tate <jess@sci.utah.edu>
To:
Cc: scirun-users@sci.utah.edu; Nima_Maftoon@MEEI.HARVARD.EDU; n_maftoon@yahoo.com
Sent: Tuesday, June 9, 2015 1:52 PM
Subject: Re: [SCIRUN-USERS] Questions about BioMesh3D

Hi Nima,

here are some answers to your question:

1)  The output of BioMesh3D is a scirun field file (.fld)  This is the native format for SCIRun, with which Biomesh is distributed.  You can use SCIRun to convert the fld file into other formats.  vtk and stl formats are supported as are several others.  Please refer to the SCIRun website (scirun.org) and documentation for more information.  

2)  There are several intermediate steps that save out files.  These are left so that you can rerun BioMesh from later stages if desired.  Also, you can use them to take advantage of the interactive mode  and visualize the results of the various stages.  The surface file is saved out as particle-union.ts.fld.  You will need to transform it back into the original coordinate system (can be done is scirun).  You can also use SCIRun to extract the surfaces from the tet mesh.    

3) This is not supported.

4) You can set the max volume of the tet elements with the tetgen flags in the model_config.py file (see http://wias-berlin.de/software/tetgen/).  You can also play with the sizing field variables in the config file.  

Hope that helps,
Jess






On Jun 9, 2015, at 11:26 AM, Elizabeth Jurrus <liz@sci.utah.edu> wrote:

A few questions for the mailing list...
- liz


-------- Forwarded Message --------
Subject: Questions about BioMesh3D
Date: Tue, 9 Jun 2015 15:58:08 +0000
From: Maftoon, Nima <Nima_Maftoon@MEEI.HARVARD.EDU>
To: liz@sci.utah.edu <liz@sci.utah.edu>

I have few quick questions. I played with the Mickey example and I think I got the ideas to some degree. What I don't understand yet are the followings:

1) Is the final mesh file, the binary .fld file in the "junctions" directory? What kind of format does it have? Which software reads this mesh format? Is it possible to have the mesh in a more general format like .unv, vtk or even stl format to be more accessible? med is a good new format as well (http://www.salome-platform.org/user-section/about/med).

2) I see that some intermediate node, face, ele and pts files are generated. What are these? If I just need the surface mesh information which files can give me that? A user may prefer to generate my volume elements myself in another meshing software.

3) It is said that 3D tetrahedral elements are generated. Is it also possible to generate 3D shell or plate elements for thin structures like membranes?

4) Is it possible to control the element size in the mesher?

I really appreciate your answers to these questions.

Nima Maftoon, PhD

Massachusetts Eye and Ear Infirmary
Harvard Medical School
243 Charles St
Boston, MA 02114
617 981 2638










Archive powered by MHonArc 2.6.18.

Top of page