Skip navigation
Currently Being Moderated
0

Special Earth Science Department Seminar

Posted by C Lowry on Sep 4, 2012 3:45:05 PM

Special Earth Science Department Seminar

 

Speculations on Two Open Problems in Morphometrics:

Random Walks and Biomechanical Strain Analysis

 

Prof. Fred L. Bookstein*

 

Department of Anthropology

University of Vienna, Vienna, Austria

 

Department of Statistics

University of Washington, Seattle, WA, USA

 

 

20 September 2012, 16:00

Neil Chalmers Seminar Room, The Natural History Museum, Cromwell Road, London

 

 

The example of morphometrics shows us that a continual refocusing of the empirical context of a method

may sometimes be a creative intellectual act, extending and clarifying the range of domains to which the

biometric imagination can contribute. In this presentation, I sketch two little-explored arcs of morphometric’s

conceptual boundary along which it may be just the right time for a collective exploration of new possibilities

with new colleagues.

 

1. Random Walk as a Null Model for Morphometric Time-Series (Paleobiology, in press)

The more interesting the process that creates your principal components, the less likely that their scatters will

look like the ellipsoids the Gaussian model has induced you to expect. Since evolution is our most interesting

biological process, principal components may be particularly ripe for an alternative interpretation when

applied to time series of fossil forms.

 

2. Form and Function in the 21st Century: The Relation Between Geometric Morphometrics and Finite-

Element Analysis (Biological Theory, in press)

Material deformations shatter the overwhelming formal symmetries of the Procrustes approach to biometric

shape comparison. Hence principal components of Procrustes shape coordinates are not an automatic

machine for the understanding of biomechanical strain or finite-element analysis. This presentation will

diagnose the problem and offer a possibly helpful workaround.

 

 

 

For additional details on attending this or other seminars see http://www.nhm.ac.uk/research-curation/seminars-events/index.html

 

* Biometer, statistical scientist and applied mathematician, F. L. Bookstein, along with important contributions by David

Kendall, Colin Goodall, Kanti Mardia and others, was the principal creator of geometric morphometrics, a new specialty

that combines the disciplines of geometry, computer science, and mathematical biology with multivariate statistics to

create tools and techniques to support the analysis of biological shape differences. Bookstein’s many conceptual and

computational innovations are being applied broadly today across evolutionary and developmental biology, paleontology,

anthropology, computer vision, medical imaging, and cognitive neuroimaging.

Comments (0)

Bookmarked by (0)