Hi all,
The answer to Pablo Jarrin's question about how to rotate a set of landmark 
configurations is simple; all you want to do is "rotate" the coordinate set 90 
degrees 
prior to computing symmetrized shapes in BigFix.  As Oyvind suggested this can 
be 
done easily in Excel.  Let's say you want  Ynew = Xold, Xnew = -Yold (up 
becomes 
left and down becomes right):  
1. Open your tps format file (from TPSdig or CoordGen) in Word, delete any 
spaces 
in control lines (lm=... etc.).  
2. Save and open in Excel.  
3. Copy the first column (Xold and control lines) into the third column.  
4. Multiply the second column (Yold) by -1.  
5. Cut and paste the control lines (LM= etc) so they start in the second column 
and 
read correctly.
6. Delete the first column (Xold).  The new first column (Xnew) is the negative 
of the 
original second column (-Yold), and the new second column (Ynew) is the 
original 
first column (Xold).
7. Save as a text file.

To rotate in the opposite direction, Up to right, all is the same except step 
4. In the 
new step 4, you multiply the third column by -1, not the second column.  Now 
the 
result is Xnew = Yold, Ynew= -Xold.
This will allow you to reflect and average in BigFix (or you could do that in 
Excel, 
also).  

With regard to the question of visualization, yes, the correct visualization of 
analyses based on the symmetrized half-skull, is that half skull.  The problem 
is that 
Procrustes superimposition treats the midline points the same as any other, and 
as 
Bookstein once pointed out "happily" rotates the shapes.  This sometimes 
produces 
disconcerting pictures in which the midline appears to rotate and deformations 
appear to produce asymmetry.  Generation of symmetrical configurations is one 
way to avoid these visualization issues, by mirroring the configuration of 
lateral 
landmarks.  The result is superimpositions that slide, but don't rotate.  It is 
important 
to understand that these pictures are strictly for visualization purposes; if 
they don't 
help you to understand and illustrate the results of analyses on the half 
skull, don't 
use them.

 

----------
Don Swiderski
University of Michigan
ph: 734-763-9613
fax: 734-615-8111
[EMAIL PROTECTED]
-- 
Replies will be sent to the list.
For more information visit http://www.morphometrics.org

Reply via email to