View Full Version : Re: Converting CT or MRI scans into 3D solid models

07-02-2004, 05:00 PM
This question appears recurrently; I guess we should find the time to
make a FAQ or something like this also for BIOMCH-L. So, the first
answer is ... look into the archives. The second is look for this
review article where most methods are described in great detail:

Viceconti, M. and F. Taddei (2003). "Automatic generation of finite
element meshes from computed tomography data." Crit Rev Biomed Eng
31(1-2): 27-72.

There are three strategies to create an FE mesh from CT data.

Direct generation
Some algorithms let you create a FE mesh directly from the CT data.
these are all so-called voxel meshes, which directly convert CT
voxels into a cartesian mesh. there are also algorithms that start
from an average template mesh for that organ and perform an elastic
deformation of the mesh onto the CT data. Non of these algorithms is
commercially available as far as I know. there is a software
developed and commercialised by some former students of Phil Young (I
cannot recall the name of the company right now) which should do
something similar, but I never tried myself.

Segmentation and meshing
This is a two steps method. Firstly you segment you organ in the CT
images, then you use the segmentation surface to create the mesh.
There are a number of commercial programs that let you segment images
and volumes. I mention Amira, which is probably the most complete,
and Mimics, the software of the Belgian Materialise which is specific
for this purpose. Usually these software provide you with a stack of
contours or a polygonal surface, i.e. a surface made of triangles.
Then the problem is to find and automatic mesh generator that is able
to cope with such surface, that is frequently full of gaps and other
imperfections. We used for a long while a software called Hexar,
which is not available anymore. We also tested another couple of
commercial software without great success. the best guess would be
the block-partitioning grid projection meshers, such as TrueGrid,
which project their grids into the surface.

Segmenting, converting and meshing
This is the longest but most reliable methods we found so far. After
you have a polygonal surface you convert it into mathematical surface
i.e. a set NURBS patches, using reverse engineering software. We use
Raindrop Geomagic, but there are others around. then you can import
the math surfaces in any preprocessing environment and mesh using



>I am trying to develop a 3D head model to perform finite element
>analysis. I created a model using the 2D sections of cadavers (using
>McGrath/Mills 1984), but it does not look good. The reason is that
>the sections are too thick. I decided to use a 3D converter software
>to convert MRI or CT scans into the solid model. Searching the web
>gave me some insights about other researches, but my problem has not
>resolved yet.
>Is there anyone who has the experience of creating solid model from
>MRI or CT scans?
>Any help would be greatly appreciated.
>Mohamad Zoghi
>Mohamad Zoghi-Moghadam
>PhD Candidate
>Department of Mechanical Engineering
>City College of the City University of New York
>140th and Convent Ave.
>New York, NY 10031
>To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
>For information and archives: http://isb.ri.ccf.org/biomch-l
>Please consider posting your message to the Biomch-L Web-based
>Discussion Forum: http://movement-analysis.com/biomch_l

MARCO VICECONTI, PhD (viceconti@tecno.ior.it)
Laboratorio di Tecnologia Medica tel. 39-051-6366865
Istituti Ortopedici Rizzoli fax. 39-051-6366863
via di barbiano 1/10, 40136 - Bologna, Italy

Tiger! Tiger! Burning bright in the forest of the night,
what immortal hand or eye could frame thy fearful symmetry?
Opinions expressed here do not necessarily reflect those of my employer

To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l
Please consider posting your message to the Biomch-L Web-based
Discussion Forum: http://movement-analysis.com/biomch_l