-------- Original Message --------
Subject: Re: Resistant Fit Procrustes Superimposition
Date: Thu, 2 Jun 2011 10:34:44 -0400
From: Greiner, Thomas <[email protected]>
To: [email protected]
I too am interested in resistant fit programs that may be out there.
I don't think the distance of the one landmark is really important in
the superimposition process, but the variation in its location will be.
So, if I understand the situation, creating pseudo-landmarks around the
isolated landmark would only magnify its potential Pinocchio effect.
Have you tried excluding your one landmark during the superimposition
process? This would have a reverse Pinocchio effect in that none of the
variation associated with this landmark will be assessed. It therefore
may not be a good solution for your problem, but it is the best idea I
can come up with at the moment.
*Thomas M. Greiner, Ph.D*.
Anatomist & Physical Anthropologist
Dept. of Health Professions
University of Wisconsin-La Crosse
La Crosse, WI 54601 USA
608-785-8476
[email protected] <mailto:[email protected]>
On Thu, Jun 2, 2011 at 8:54 AM, morphmet
<[email protected]
<mailto:[email protected]>> wrote:
-------- Original Message --------
Subject: Resistant Fit Procrustes Superimposition
Date: Wed, 1 Jun 2011 01:21:37 -0400
From: Ruth Flatscher <[email protected]
<mailto:[email protected]>>
To: [email protected] <mailto:[email protected]>
Dear all,
I am a botanist working with an object whose landmark distribution is
not ideal, i.e. one single landmark is quite far away from the others.
In addition, based on my observations I assume that this same landmark
is also more variable in its position to the others. Performing an
"ordinary" Least Squares Procrustes fit, I fear I am facing a
"Pinocchio
effect". Therefore, I have thought of using Resistant Fit Procrustes
Superimposition. Does anybody of you have experience in it, and is
there
any software available which is able to calculate it?
I have also thought of other solutions, like a Procrustes fit omitting
size standardization. Alternatively, I could calculate two
pseudolandmarks on either side of this distant landmark, in order to
make it less isolated from the others. What do you think about these
possibilities?
I would be very glad if somebody could help me out with some
suggestions.
Best wishes, and greetings from Vienna,
Ruth
Mag. Ruth Flatscher
Department of Evolutionary and Systematic Botany
Rennweg 14
A-1030 Vienna
email: [email protected] <mailto:[email protected]>
<mailto:[email protected] <mailto:[email protected]>>
