300569 VO Special topic in morphometrics (2006S)

Much of descriptive biological science talks about
organismal form, and yet comparative anatomy and
development remain peculiarly _in_formal, _un_quantitative
relict fields within 21st-century
biology. Over the last hundred years there have been
a variety of attempts to make these studies objective;
none are obvious successes. In this short course I will discuss a
range of these strategies, emphasizing the different rhetorics
that have been tried out (physical, cognitive, geometrical, statistical).`
I will conclude with an exposition of my own currently favorite
possibility, the separation of shape factors into global vs.
local, to be treated separately by quite distinctive methodologies
unified only at the level of their formal geometry.

Sketch of topics (selection subject to student preferences):

Prehistory: portrait art and caricature

At the turn of the twentieth century: anthropometrics
and odontometrics

Shape from physics: D'Arcy Thompson and explanation by analogy

Multivariate morphometrics and its failure
shape distance -- not as useful as we thought

Two crucial distinctions:
Shape as biometric variable/Shape as explanation
global versus local

Versions of global
uniform terms
growth-gradients
thin plates and "energy"

Versions of local
characters
medial decompositions
radial basis functions
edge features in TPS

A radical suggestion
intrinsic random fields: rationale and example(s)