Generalizing and extending the eigenshape method of shape space visualization and analysis

Abstract

Outline-based morphometric methods have been more or less restricted to the consideration of closed curves and plagued by problems related to the maintenance of close biological correspondence across all forms within a sample. Methods developed herein generalize and extend the eigenshape method of outline analysis along the following lines: (1) consideration of open curves, (2) improvement of interobject correspondence via incorporation of information provided by landmarks, and (3) extension to the analysis of three-dimensional (open and closed) curves. In addition, techniques for using eigenshape results to create models of shape variation and for more consistently assessing the digital resolution necessary to represent an object are discussed and illustrated. These improvements are then placed in context via discussions of previous attempts to extend morphometric outline analysis methods, the relation between landmark and outline-based morphometric methods, the use of morphometric analyses to test biological hypotheses, and the nature of morphometric shape spaces (with special reference to studies of morphological disparity).