[Sorry for the delay in posting this and the previous message, but I was
unexpectedly trapped out of town and away from the internet for nearly a
week. - the Moderator (dslice)]
Many curve digitization packages (e.g., tpsDIG, ImagePro Plus) give you the
ability to sample curves with a fixed number of equally spaced (or
quasi-equally spaced) semi-landmark points. The x,y -> ext. phi program in
my extended eigenshape package indeed assumes you have collected your data
in this format. I can think of a few obvious workarounds and, if there is a
need for the capability of reading in outline strings of different lengths
in that program, I'd be happy to add it. Given the capabilities of tpsDIG,
however, it would seem the best thing to do is collect the data in the
expected format to being with.
Please contact me offline if you need instructions has to how to make tpsDIG
collect outline strings of equal lengths.
Norm MacLeod
___________________________________________________________________
Prof. Norman MacLeod
Keeper of Palaeontology
The Natural History Museum
Cromwell Road, London, SW7 5BD
(0)207 942-5204 (Office)
(0)207 942-5546 (Fax)
http://www.nhm.ac.uk/palaeontology/a&ss/nm/nm.html (Web Page)
___________________________________________________________________
On 23/2/06 20:24, "morphmet" <[EMAIL PROTECTED]> wrote:
Dear All,
My name is Gordon Smith, a PhD student at Sheffield University.
my e-mail is [EMAIL PROTECTED]
My query is about open outlines and extended eigenshape analysis.
I have what amounts to 2D open outlines, for a number of individuals, as
X,Y coordinate data. After my minimal literature review, I thought that
the best way to analyse this data would be with extended eigenshape
analysis using the Norman Macleod software from the Stony Brook website.
However, after playing with the software and further reading it seems that
the number of X,Y coordinates for each open outline should be the same for
every individual in the comparison.
This is not the case for my dataset. I have found a number of software
packages that appear would calculate an equal number of X,Y coordinates
per open outline from image data such as bitmaps, jpegs, etc. but not from
collected X,Y data.
So I have two questions really, Firstly is my understanding correct and
that to be able to compare individuals will they all require an equal
number of X,Y coordinates along the open outline.?
Secondly, If this is correct can anybody suggest how I make my current X,Y
coordinate data compatible with the analysis.?
I am very new to the world of morphometrics, and would welcome any and all
help,
as to key papers, books etc.
Please reply if any further information would be helpful to answer this
query.
Many Thanks
Gordon.
e-mail [EMAIL PROTECTED]
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
