Measurement errors in 3D models used in osteometric data research with freeware: a test using skulls of the algerian hedgehog (Atelerix algirus)

Abstract

The use of 3D models is becoming commonplace in both scientific investigation and in more general educational activities; nevertheless, methodologies and their associated software are both relatively expensive to use. As well, measurements of 3D models must be comparable to that of originals if they are to be used in scientific research. The aim of this study was to determine the degree of error in measurements taken from 3D models and from the original skulls of Algerian hedgehogs (Atelerix algirus) using a freeware program. To do so, we measured the repeatability (ri) of four biometric measurements of 14 skulls of this North-African hedgehog species. We generated 3D models of skulls of 14 specimens from the collection in Barcelona Natural History Museum using the NextEngine scanner and free software and a low-price computer. The same observer measured each variable twice for each skull using three different methodologies: (i) measurement of the original; (ii) measurement of 3D models projected onto a screen with no zoom (i.e. replicating the original skull size) and (iii) measurement of 3D models projected onto a screen with the maximum possible zoom. The repeatability within each method (INTRAMETHOD) and between methods (INTERMETHOD: original vs. screen No-zoom; original vs. screen zoom) were tested. The methods per se were either very highly repeatable (ri > 90%) or very repeatable (ri > 0.75), the only exception being a difference in the length of the median palatine on the screen with no zoom (ri: 0.64) due to a single outlier. When comparing digital models, our data suggest that measurements taken from skull borders are more reliable than those in which measurements are taken from sutures due to the differences in the contrast obtained in the finish of the 3D models. Thus, the contrast in 3D models needs to be improved, possibly by means of digital treatment. Our results suggest that the 3D models obtained using the scanner NextEngine and edited using open-access freeware (Meshlab®) are comparable with original specimens and so are a good alternative for museums with fewer financial resources.