View Full Version : Summary of Kinemetrix

Dr David Michael Hooper
09-25-1996, 08:22 PM
Hello Biomch-L readers,
I recieved a number of responses to my initial posting
regarding camera and marker configurations to be implemented
during a study of gait using a three camera Kinemetrix system.
The original posting is included following this summary.
So, this is basically what I have come up with....

1. Several individuals suggested placing a non reflective screen
between the subject=92s legs so that the side cameras would be shielded
from the opposite leg. This idea may be effective, but I=92m not
certain that the presence of the shield would not effect the subjects

2. Dr. Lang Yang from Noth General Hospital in Sheffield, England
informed me that he had used masking tape to cover a potion of the
marker to hide it from the opposite cameras. As the markers are
spherical, this method may work but I believe that this method will be
difficult during this dynamic task. Also, the tape will alter the
center of reflection seen by the middle camera, and may induce an

3. These two concepts, gave me an idea which has worked although now I
have other problems related to the software. Anyway, I found that by
placing double sided tape on a small piece of paper (i.e. a
post-it note) I could stick the paper to one side of the sphere. The
paper acts to shield the view of the opposite camera, while allowing
easy detection of the entire marker by the two appropriate cameras. I
implemented this in the lab yesterday, and it works well.

Thank you for the suggestions. And again, if anyone is working with
this system in particular, I would invite you to contact me to
exchange knowledge on it=92s implementation.

Dr. David M. Hooper
University of East London
Tel: 0181 590 7000 ext. 4025

************************************************** ****************
********************* ORIGINAL POSTING ****************
************************************************** ****************
I have recently taken a position at the University of
East London and am being charged with bringing a motion
analysis laboratory into full swing. My question (or query
in the UK) concerns the collection of data using a Kinemetrix
camera system and infrared reflective markers. My dissertation
work involved the Selspot II system which utilizes infrared LEDs,
active versus passive markers and I am encountering some problems
which I am am unfamiliar with.
I am trying to construct an experiment to look simulataneously at
both legs during gait. My idea is to have markers on the back of the
lower limb segments and the three cameras located symmetrically
behind the subject, who is walking upon a treadmill. Since each
marker must be seen by at least two cameras at each instant in time, I
had intended to have a central camera looking at all markers...while
cameras on the left and right sides looked at markers on the left and
right sides, respectively. The problem arises when the markers of the
left leg are intermittently visible to the camera on the right side
(due to interference from the right leg). Since the system is passive
it has no way of knowing which marker is which, and this is causing
problems in reconstructing the 3d coordinates appropriately.
Does anyone know of a way to clean up the individual cameral views
*then* use those cleaned views to construct 3d coordinates. As far as
I am concerned, I just want to be able to =93erase=94 the markers on the
left leg from the view of the camera on the right, and similarly on
the left.
If anyone has encountered this problem, or would like to bounce
concerning the Kinemetrix system, please contact me. I would be quite
interested in discussing this with you. Naturally, I will post a
summary of replies.

Thank you.

David Hooper
Department of Health Sciences
University of East London
************************************************** ************************=