Standard methods for creating digital skeletal models using structure-from-motion photogrammetry.

Area of Science:

• Anthropology
• Digital Imaging
• Biomedical Engineering

Background:

• 3D digital modeling of skeletal remains is crucial for research and preservation.
• Structure-from-motion (SfM) photogrammetry offers a non-invasive method for creating these models.
• Standardized methods are needed to ensure the accuracy and comparability of photogrammetric models to physical specimens.

Purpose of the Study:

• To assess best practices for producing 3D digital cranial models using SfM photogrammetry.
• To determine if photogrammetric models are metrically accurate and visually comparable to physical crania.
• To establish a user-friendly standard method for creating accurate digital skeletal models with Agisoft PhotoScan.

Main Methods:

• Approximately 200 photographs of three crania were captured and divided into series (50-200 photos).
• Forty-five cranial models were generated using various photo series and Agisoft PhotoScan settings.
• Model accuracy was assessed by comparing measurements to physical crania using Bland-Altman plots and qualitative criteria.

Main Results:

• The majority of models (37/45) showed high agreement with physical measurements (mean difference ≤ 2 mm).
• Models created with 150 photographs and "high" PhotoScan settings achieved the best qualitative scores in the shortest processing time.
• Photogrammetric models demonstrated comparable metric accuracy and overall presentation to physical crania.

Conclusions:

• Creating 3D digital cranial models with 150 photographs and "high" settings in Agisoft PhotoScan is recommended for metrically accurate and visually comparable results.
• SfM photogrammetry is a versatile, rapid, and non-invasive tool for generating accurate 3D digital models in various settings.
• Adhering to the established guidelines ensures the metric accuracy of SfM-derived digital cranial models.