-------- Original Message --------
Subject: Re: Comparing virtual bivalve shell morphologies
Date: Fri, 13 Feb 2009 23:13:37 -0800 (PST)
From: Brendan McCane <[email protected]>
Reply-To: [email protected]
To: [email protected]
References: <[email protected]>
Dear Daniel,
I am not familiar with the work of Norm, but another alternative might
be to look at the Iterative Closest Point (ICP) algorithm which can be
used to align meshes. The reference is:
@article{besl1992mrd,
title={{A method for registration of 3-D shapes}},
author={Besl, P.J. and McKay, N.D.},
journal={IEEE Transactions on Pattern Analysis and Machine Intelligence},
volume={14},
number={2},
pages={239--256},
year={1992}
}
An algorithm could then work as follows:
start with an estimate of the mean shape (choose a sample)
while mean hasn't convergeg
for each sample align with mean using ICP
recalculate mean
Extract semi-landmarks from the aligned set and use standard GPA on the
result.
morphmet wrote:
-------- Original Message --------
Subject: Re: Comparing virtual bivalve shell morphologies
Date: Fri, 13 Feb 2009 00:28:04 -0800 (PST)
From: Norman MacLeod <[email protected]>
To: Morphmet List <[email protected]>
You might want to check out the description of eigensurface analysis in:
MacLeod, N., 2008, Understanding morphology in systematic contexts: 3D
specimen ordination and 3D specimen recognition, in Wheeler, Q., ed.,
The New Taxonomy: London, CRC Press, Taylor & Francis Group, p. 143210.
It does more-or-less exactly what you describe in a standard geometric
morphometric way.
Norm MacLeod
On 12/2/09 18:21, "morphmet" <[email protected]> wrote:
-------- Original Message --------
Subject: Comparing virtual bivalve shell morphologies
Date: Thu, 12 Feb 2009 03:04:20 -0800 (PST)
From: Daniel Germann <[email protected]>
To: [email protected]
Hi all
In our project, we are investigating the functional morphology of
bivalves for burrowing. We will work with 3D computer models of bivalve
shells both resulting from CT scans of real shells and generated using
mathematical growth models. So, our data consists of 3D points or
polygon meshes representing the geometry of the shells.
The important question is now how to measure similarity between two such
virtual shell models. Probably one has to extract morphological traits,
quantify them and then compare these values. The problem is that bivalve
shells do not have many landmarks and that we need to compute the
similarity between two meshes automatically, i.e. whithout any human
interaction like marking interesting points. For artificially generated
shells, we have parameters like the position of the umbo, the shape of
the generating helicospiral or the aperture curve. It would be nice to
have a method to compute this also for scanned shells.
Does anybody know a good method or even an available software to
quantify the above morphological traits (umbo position, helicospiral,
aperture curve etc.) for bivalve shell models? What are the important
morphometric quantities for a good similarity measure? Any hints are
highly appreciated.
Thanks a lot!
Best regards
Daniel Germann
-----------------------------------
Daniel Germann, MSc ETH CS
Ph.D. Student
Artificial Intelligence Laboratory
Department of Informatics
University of Zurich
Andreasstrasse 15
CH-8050 Zurich, Switzerland
Phone: +41 44 635 45 09
Email: [email protected]
-----------------------------------
___________________________________________________________________
Prof. Norman MacLeod
Keeper of Palaeontology
The Natural History Museum
Cromwell Road, London, SW7 5BD
(0)207 942-5204/5295 (Office)
(0)207 942-5546 (Fax)
http://www.nhm.ac.uk/hosted_sites/paleonet/MacLeod/ (Web Page)
___________________________________________________________________
--
Cheers,
Brendan.
----------------------------------------------------------------------------
Brendan McCane, Head of Dept, Email: [email protected]
Department of Computer Science Phone: +64 3 479 8588/8578.
University of Otago Fax: +64 3 479 8529
Box 56, Dunedin, New Zealand.
--
Replies will be sent to the list.
For more information visit http://www.morphometrics.org
