32 Photogrammetry 3D Models : A Novel Approach to Surgical Medical Education but What About Their Accuracy?

Surgical teaching and learning of anatomy are mostly based on the donation of human bodies. It is acknowledged that students learn anatomical structures best by dissecting cadaveric specimens or the exploring prosected specimens. However, the availability and accessibility of these specimens is limited due to their confinement to licensed facilities and their short-shelf life. The integration of 3D digitization and modelling in medical education has delivered promising results. Despite the progress in 3D imaging techniques, the accuracy and reproducibility of these models have not been thoroughly investigated. This paper seeks to show the various factors which may alter model inaccuracy and irreproducibility. Six archaeological bone human specimens were digitized in this study. The photogrammetry models were compared to micro-CT-based reference models and a low-cost 3D laser scan. Results show that photogrammetry provides overall the most accurate reconstruction of the specimens across all the comparisons. The inter-and intra-observer error was also assessed. Results showed no significant trend in intra-observer variability. In regard to inter-observer variability, a minimal variation was recorded. This is significant as it demonstrates that the data obtained is reproducible and robust, in particular applicable for the repeatability of the same data from different users with minimal overall variability. This paper has highlighted the accuracy of photogrammetry as a useful low-cost tool for the 3D reconstruction of cadaveric models for surgical medical education.