View Full Version : Re: Surface markers in 3D analysis

Scott Z. Barnes
12-09-1995, 10:07 AM
In regards to the ongoing discussion, Dr. Andriacchi presented an
applicable method during a talk at CSB 1994 in Calgary entitled "New
Developments in Musculoskeletal Functional Testing".

The technique he presented was the Point Cluster Technique (PCT) and in
brief, a portion of the abstract reads...

"The point cluster technique is based on a cluster of points distrubuted on
the limb segment. Each point is asigned a unit mass. The center of mass
and the inertial tensor of the cluster points is calculated. The
eigenvalues and eigenvectors of the inertia tensor are the principal
moments of inertial and principal axes of the point cluster. The
eigenvalues establish a transformation between the segment and the clobal
coordinate system"

and so on.... I think this is enough to get the idea. Dr. Andriacchi went
on to point out that the cluster of points (markers) could be placed in
random fashion and that the problems related to marker motion, skin motion,
and other sources of error are minimized by the technique.

The method was also validated by utilizing two test subjects with Lizarov
devices (the stainless steel exo-skelaton used in leg lengthening, etc.) I
wonder how they got that past the human subjects approval committee?? (just
kidding!, there happend to be two willing patients available at the time).
This allowed the procedure of using markers on a subject's soft tissue
(similar to typical gait analysis), to be compared to bone fixed markers
(various points on the Lizarov).

The results were very encouraging and accurate. Perhapsse Dr. Andriacchi,
or some others in his research group, care to comment on new developments

A couple other points for discussion.

1. I think the basis of this method, point cluster and eigen solutions,
could have some relavance regarding the standardization of coordinate
systems and data sharing being undertaken. Similar methods could be used
in order to transform data taken at different labs or from different
systems (motion tracking hardware) into more compatable forms for
comparison. Similar solution techniques are the heart of computer graphic
solutions that scale a generic data base (skeleton surface model for
example) to specific subject data. Thus allowing for patient or subject
specific graphical analysis of data without having to perform a complex and
costly imaging scan of some type.

2. The number of points that need to be tracked for whole body studies are
obviously increased, in most cases, beyond the capabilities of existing
hardware/software combinations. I do not think there are commercial
systems with this capability built into the analysis modules. I think Dr.
Andriacchi's group utilized SGI's and custom software??

3. This could be a good way to attack small volume or zoomed volume
problems such as hand and wrist, foot specific, or joint specific, where
increased accuracy is required.

Just some thoughts.


Scott Z. Barnes
The Ohio State University
H: (614) 442-7797
W: (614) 421-2803
F: (614) 421-2803