Detailed Anatomical Orientations for Certain Types of Morphometric Measurements Can Be Determined Automatically With Geometric Algorithms.

Abstract

Morphometric datasets only convey useful information about variation when measurement landmarks and relevant anatomical axes are clearly defined. We propose that anatomical axes of 3D digital models of bones can be standardized prior to measurement using an algorithm that automatically finds a universal geometric alignment among sampled bones. As a case study, we use teeth of "prosimian" primates. In this sample, equivalent occlusal planes are determined automatically using the R-package auto3dgm. The area of projection into the occlusal plane for each tooth is the measurement of interest. This area is used in computation of a shape metric called relief index (RFI), the natural log of the square root of crown area divided by the square root of occlusal plane projection area. We compare mean and variance parameters of area and RFI values computed from these automatically orientated tooth models with values computed from manually orientated tooth models. According to our results, the manual and automated approaches yield extremely similar mean and variance parameters. The only differences that plausibly modify interpretations of biological meaning slightly favor the automated treatment because a greater proportion of differences among subsamples in the automated treatment are correlated with dietary differences. We conclude that-at least for dental topographic metrics-automated alignment recovers a variance pattern that has meaning similar to previously published datasets based on manual data collection. Therefore, future applications of dental topography can take advantage of automatic alignment to increase objectivity and repeatability.