STRUCTURE FROM MOTION AS A TOOL FOR 3D MORPHOMETRICS: A NEW PALEOBIOLOGICAL APPLICATION FOR A YOUNG GEOLOGICAL METHOD

Morphometrics is an invaluable tool for quantifying the shape of fossil organisms. It can be used to determine the extent to which individuals are similar and different, which can then be correlated to evolutionary, ecological, or developmental changes. The simplest and most accessible visualization methods (e.g., landmarks, outlines, etc.) perform best in 2-dimensions, which means there is limited ability to quantify shape in 3-dimensions. This is problematic, especially when trying to better understand the relationship between form and ecological function, because many morphological characters thought to be important to this relationship (e.g., globosity, ribbing, texture) are expressed in 3-dimensions. To robustly quantify these shape parameters, it is important to employ a 3D morphometric technique. We argue that a recently developed visualization method called ‘structure from motion’ (SfM) is ideal for such a task.

SfM is a method of creating a 3-dimensional image from 2-dimensional photographs of an object. This involves taking a series of overlapping photographs with any digital camera (even a camera phone). Photographs are then imported into a computer program that aligns them and converts them into a 3D digital model, essentially a cloud of points. Point clouds from multiple objects can be exported as a data file of points and analyzed via an ordination method, e.g., principal component analysis (PCA).

The application of SfM in the geosciences has so far been mostly imaging outcrops, but its utility is slowly being expanded to hand samples, suggesting that it has great potential for 3D morphometrics. We have developed a method for using SfM to create 3-dimensional point clouds of brachiopod shells. Here we present best practices learned from method development and some preliminary results of Ordovician brachiopods.

SfM is advantageous because it is inexpensive, easy to use, and produces high-resolution images. Its low cost and simplicity is preferable to more technologically complex 3D methods and makes generating large morphological data sets for robust statistical analysis possible. Its ability to capture 3-dimensional features allows for a more complete analysis of shell shape. These advantages make SfM ideal for any fossil organism and an important development to the field of morphometrics.