View Full Version : Geometric models

Rhona Phelps
07-13-1997, 07:58 PM
Dear All

A couple of weeks ago I asked a question about 3D model creation.
Basically, I wanted to know if there were people out there creating
models from anatomic data and if so how were they doing it and would
they be prepared to share their models.

A big thanks to those people who did reply.

I recieved a number of messages asking me to send them a summary of
my replies - the question itself turned out to be very common, but
the answer less so. The following are the responses I recieved. I
would welcome any further replies.



>With regard to your questions regarding the VHP, I followed the
discussion last year but found most people unwilling to divulge their
hard work, especially those that had developed models for use in
CAD/FE. It does seem silly not sharing this info as it is only a
small aspect of work and could be of great use to a large number of
people. I feel sure that somebody must have published this as part of
their thesis which is probably very difficult to track down. The
commercial software seems a bit expensive.
I'm afraid that at this moment I do not have any models to share,
what are you looking for in particular?? We have developed models of
the knee, spine and ankle by a technique involving physical
measurement using a co-ordinate measuring machine and IGES files.. I
am currently working on a technique to develop models from the VHP
dtaa, however this work is not complete yet. The method I intend to
build involves using the CT data and putting it into IGES format.
Dont know to what extent all of this will be done automatically, it
may require some input and manipulation.
Chris Connor, email: C.J.Connor@tees.ac.uk
School of Science & Technology University of Teesside

>I am sending you a copy of a technical report that describes the
spatial geometry of the human pelvis. I conducted a study with Clyde
Snow and Joe Young several years ago in which we measured 165
skeletal pelves from the Cleveland Museum of Natural History. The
results were used to develop small female, average male and large
male pelves for the US Department of Transportation. The
three-dimensional data describe the location of approximately 122
points on the pelvis. I believe that you can use this information
in your model quite easily. If you have any questions after receiving
the report, please feel free to contact me.
Mac Reynolds, Ph.D. , email: reynolds@ergo.msu.edu
Ergonomics Research Laboratory Michigan State University

>In reply to your request for numerical data of the pelvis and femur,
I can tell you that I am working on a mathematical model of the
pelvis/low back. The aim is to calculate muscle forces, forces in
ligaments and joint reaction forces in several postures and under
several loads. (So it will NOT be a FEM of bony structures). For this
model, I need lots of geometrical data, so we made a set of
MRI-scans of one subject. The contour of all structures that are
important for me was determined by hand: drawing contours with the
computer mouse. The final result is a data set of contours of -among
others- the pelvis and the femur. The distance between the slices is
5.5 mm, the accuracy of the contours is determined by the skill of
the person that draws the contours, I can't give a quantitative
measure for that. These contoures are stored in a numerical format
in my computer.
Gilbert A. Hoek van Dijke, email: hoekvandijke@bnt.fgg.eur.nl
Department of Biomedical Physics and Technology Erasmus University

>In recreating the geometry of soft tissue knee structures from
magnetic resonance images, we have been using three pieces of
software, PV Wave, Nuage, and GeomVIEW. I would attempt to give you
more detail, but I am learning myself. I am currently taking a
class in AutoCAD in hopes of replacing or aiding the programs which
we use now.
Tony Francisco, email: acf@acpub.duke.edu
Duke University Sports Medicine

>Some years ago I made efforts to analytically represent the geometry
(asphericity) of the human femoral head. My results were published in
two South African Journals ( my post-doc was in this country):
-Mizrahi J: The human femoral head as a tilted solid of revolution.
South African Journal of Science, 73(9), 385-386, 1977. -Mizrahi J:
An axi-symmetrical representation of the human femoral head. The
South African Mechanical Engineer, 28, 206-209, 1978.
Prof. Joseph Mizrahi , D.Sc., email: jm@biomed.technion.ac.il
Dept. of Biomedical Eng. Technion,
Israel Institute of Technology

>You should take a look at http://www.zygote.com or
http://www.viewpoint.com. They have NURBS models of what you are
looking for. They've built them after digitizing real bones.
Richard Bastien email : richard.bastien@decathlon.fr
R&D department, Decathlon

Original Message

I am about to start the process of creating a basic 3D model of the
pelvis and proximal femur, for use in both FE and CAD analysis.

With the arrival of the Visible Human data set and the availablilty of
this data for all researchers, I know a number of centers are spending
time and resources creating geometries from the images. There are
also comercial companies using this data, but the models they are
creating are 3D models in a non-mathematical format.

I searched the Biomch-l archives and came across a discussion from
last year regarding sharing of model geometry. There was however no
conclusion to this discussion. For each individual researcher to
undertake the same process of model creation from the same data would
surely be futile. Or pehaps I am simply trying tojustify my lazyness.
I could create the model myself, but as this is simply a tool for
further analysis, I was hoping the basic geometry in a numberical
format was available from some other source.

My question to the biomch-l community is, do any of you have a
goemetric model that you would be prepared to share with another
researcher? Again this should not only apply to my case, but in
general. Alternatively, could you tell me how you approached the
problem of creating the goemetry from either the scanned images, the
CT or the MRI images.

Dr Rhona Phelps Industrial Research Ltd
Research Scientist P.O.Box 2225
Engineering Dynamics Auckland
Biomechanics Group New Zealand

Email: r.phelps@irl.cri.nz
Tel: +64 9 3034116
Fax: +64 9 3070618