-------- Original Message --------
Subject: RE: Outlines in tpsDig
Date: Fri, 7 Nov 2008 09:19:20 -0800 (PST)
From: Sarah Degroot <[EMAIL PROTECTED]>
To: <[email protected]>
References: <[EMAIL PROTECTED]>
I have been able to "rig" a way to get a constant starting point for
outlines in tspDig. The program finds the outline by searching to the
left of wherever you click, and when it finds the second "edge," that's
where the points start (details are in the help documentation). I
rotated all of my images so that the one homologous point was on the
left side. Before I click to find the outline, I first note the
y-coordinate of the landmark position, then mouse to the right until the
cursor is out of the image. When I click, the first edge seen is
somewhere around the image, while the second edge is exactly on the
landmark point. Since my sample shapes are rather complex, however, I
still check the tps files to make sure the points are starting in the
right place--sounds tedious but actually not very difficult since the
x-coordinate is at its minimum on the homologous landmark.
This all does take a little longer than if I were able to click exactly
on the homologous landmark position or right next to it, but at least it
is appreciably faster than placing 220 points by hand.
________________________________
From: Sarah Degroot
Sent: Thu 23-Oct-08 11:35 AM
To: [email protected]
Subject: Outlines in tpsDig
I am trying to make my tps files multitask-- I have black & white images
and want to digitize outlines such that I can use the points in both EFA
and a sliding landmark analysis. However, some of the specimens are
fairly complex, such that I need at least 220 points to capture the
outline adequately. I don't want to place 220 points by hand,
especially since my preliminary analysis involves over 570 specimens. I
tried drawing a curve with tpsdig, but that took 28 points (minimally),
and it captured the outline poorly (even though I resampled the curve
with more points). I'm also not eager to place even 28 points on each
of 570+ specimens--there has to be a faster way.
It seems the easiest thing to do would be to use the outline tool in
tpsdig, but this tool appears to be inconsistent about where it places
the starting point. I click just to the right of where I want the
starting point, but when I look at the tps file, the starting point is
clearly somewhere else around the outline, and usually nowhere near
where I clicked. This is the case for both tpsdig 2.09 and 1.40 (1.40
has given me a few other outline anomalies, too, such as negative
coordinates).
I need the starting position of the x-y coordinates for the outline to
be in the same place on all specimens because (1) I am not correcting
for starting position in EFA because I do not want the outline rotated
(they are all oriented the same way, but the long axis is not
necessarily the same), and (2) I want one fixed "landmark" as a
reference point for the sliding landmark analysis.
I'd like something similar to Perez et al.'s (2006; J. Anat. 208:
769-784) placement of semilandmarks around molars, although I'm not
quite sure how they did it. I can't imagine they placed all 80 points
by hand (p. 772).
Suggestions are appreciated.
Sarah
Graduate Student
Rancho Santa Ana Botanic Garden / Claremont Graduate University
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
