PDA

View Full Version : summary of replies to 'bone surface mesh generating'



Fang Lin
09-15-2000, 01:41 AM
Hello, all,
Thank you for all of you who replied to my question about generating bone surface mesh. 14 colleagues replied me and provided very useful info and share valuable experience on it. Some of them even sent me their software or gave me kind offer to process my data. I didn't post this summary until I finally found the solution and sucessfully put my bone digitization data into SIMM yesterday. The original message I posted 3 weeks ago is cited here followed by all the responses.
Amanda Fang Lin, Ph.D
SMPP
Department of Physical Medicine & Rehabilitation/Rehabilitation Institute of Chicago
Northwestern University Medical School

Original message on Aug. 24, 2000:
'I am trying to generate a mesh file for bone surface from randomly digitized surface data of a bone. While I have the difficulty to find the appropriate software to automatically generate the mesh. I tried the NUAGES, someone suggested, but it needs the data on parallel z-planes, while my data is randomly distributed. So, any one can give me good suggestion? I'll forward a summary for it later.'

Responses:

Dear Fang,
We collect random points on three-dimensional anatomic surfaces via our custom build 3-D surface scanner and use 'PV Wave' software from Visual Numerics Inc. to create continuous 3-D surfaces.
This Software has a function called "FastGrid3-D" which will quickly take your surface points (in our case 64000) and will form a 3-D surface.

Michael
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Michael Bottlang, Ph.D.
Research Director
Biomechanics Laboratory
Legacy Clinical Research & Technology Center
1225 NE 2nd Ave, Portland, OR 97208

Major finite element meshing program (PATRAN, HYPERMESH, FEAMAP, etc) all support automatical surface meshing if you have the CAD surface of the bone. There are commercial surface reconstruction software allows you to generate CAD surface from digitalized surface (the one I tried was SurfDriver).
Personally and arguably, I prefer structured FE mesh to automatically generated mesh, for the reason of the quality and the control of mesh density. TrueGrid by XYZ scientific is a good software in generating structured mesh. And it support digitalized surfaces (in ASCII/binary viewpoint format as well as IGES format). You can generate high quality meshes, if you are willing to spend enough time to understand the topology of the structure to be meshed and come up with a good meshing strategy.
==
Weixin Shen, Ph.D., Senior Scientist
Science and Engineering Technology Division
JAYCOR, Inc
3394 Carmel Mountain Road
San Diego, CA 92121


Try Geomagic Sudio from Raindrop Geomagic, Inc. They're in Research Triangle Park, NC Ph (800) 251-5551 www.geomagic.com email inquiry@geomagic.com

Doug



Have a look at www.alphashapes.org . There is free software for Unix machines. For PCs or Unix, VTK (www.kitware.com ) has several surface reconstruction algorithms and is also free.

Good luck,
Goodwin

hi,
you should try "rhino3D" it's a nurb surface generator.
www.rhino3d.com

good luck,
todd doehring, m.s.
musculoskeletal research center, dept. of orthopaedic surgery
univ. of Pittsburgh

Hi Amanda
Many software programs require parallel z slices to create the surface 'mesh' and subsequent solid model. My approach would be to use least squares to fit a 3D surface to your randomly obtained data. This gives you a single equation z=f(x,y) that should accurately represent the data, assuming you have a good number of original data points which are evenly distributed.
The parallel planes you need take the form z = a where a is the distance of the plane from the origin. Thus by solving these simultaneous equations for each a, you will obtain a number of 3D curves of the form f(x,y) = 0. From these curves, data points can be resampled (ie a polar fourier series) or the curves can be used to form splines, IGES curves etc. Depends on your software.
Let me know if this needs clarifying. Good luck.
Dan Barker
Department of Orthopaedics
Lund University Hospital
Sweden
Email: dan.barker@ort.lu.se

In the past I have used a program called "Points2Polys" which is produced by a company called Paraform. It is (or was) a free download from their web-site, which is: www.paraform.com If you can't find it, I probably still have the installation files somewhere (it runs on a Windows PC, preferably with 3D accelerator card).
"Points2Polys is a high performance solution for triangulating unstructured point clouds"
As input, it takes a simple text file consisting of a series of X Y Z points. There are some examples to practice on. It can output to the following formats:
OBJ
STL Ascii
STL Binary
Paraform Files

If you need other formats, then there are other freeware tools available elsewhere to convert the files above into something you can use.
best of luck,
James.
----------------------------------------------------------------
James Ward (j.w.ward@dcs.hull.ac.uk)
Hull Medical Engineering Centre (HulMEC), Computer Science,
University of Hull, Cottingham Road, HULL (UK), HU6 7RX
----------------------------------------------------------------


Amanda,

Check out Geomagic (http://www.geomagic.com). Its sole purpose is generating surfaces from point-cloud data. We use their Shape product, which adds NURBS to their excellent polygonal Wrap program (There are a fair number of packages that can generate polygonal meshes from point-cloud data, but relatively few that will generate parametric surfaces simply). They have a substantial educational discount, and the program (Shape) is exceptionally easy to use. Runs on NT & SGI.

--Matt
************************************************** ********
Matthew P. Reed, Ph.D.
Assistant Research Scientist
University of Michigan Transportation Research Institute
************************************************** ********

Amanda,

I think Rhinoceros (http://www.rhino3d.com) will do the trick. It will convert your point cloud into a NURBS surface and can subsequently triangulate it.

Best regards,
John

Amanda:

A package called Raindrop Geomagic will do what you desire, but it is rather expensive for a commercial license. Of course, you may get a substantial discount as an academician. You can learn more at www.geomagic.com or
mailto://inquiry@geomagic.com.

Regards,
--
Anthony Petrella
Senior Research Scientist
Biomechanical Testing and Analysis
DePuy, a Johnson & Johnson company
700 Orthopaedic Dr, Warsaw, IN 46581-0988



Dear Amanda,

I saw your posting regarding the creation of mesh files. Our product, The MotionMonitor captures and synchronizes data from Ascension magnetic trackers, video, emg, forceplates and other A/D sources. One of the new features announced at the ASB conference in Chicago included a function for tracking the interaction of joint surfaces. It includes a routine for digitizing bone (or other) surfaces, functions to create mesh files and viewing functions to display wire frames or gourad shaded images of the bone interaction during activities that were collected with The MotionMonitor.
You can see a description of the procedure and an image of the mesh file at:
http://www.innsport.com/in-vitro.htm
________________________
Lee E. Johnson
ljohnson@innsport.com
Innovative Sports Training, Inc.
....The Total Solution in Motion Capture.
www.innsport.com

Have you tried the Hypermesh software?
www.altair.com
It's a very powerful meshing package but not very user-friendly.
Please post a summary,

Good luck,
Laura

_________________________
Laura Finney
Bioengineering Unit
University of Strathclyde
Glasgow, Scotland, UK

Sorry for the late of this reply. We use for this a commercial program called geomagic . It works. If you have to deal with this problem once every while it may be worth consider it. It is quite expencive but for research purpose we obtained a big rebate (still not chep, over 2000 USD).
If you have to do it just once, no easy solutions. NUAGES doesnot work for oyur problem as you found already. If you can tell me more about your project maybe we can collaborate somehow, and I could process the data for you ....
We have also very sophisticated meshing facilities so it may be of interest.
Please look at our web site: http://www.ior.it/tecno/ under computational
biomechanics
Ciao
Marco

---------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
---------------------------------------------------------------