Evaluation of different technologies used for extraoral surface data acquisition for 3D facial scanning

Purpose

This study evaluated the accuracy (trueness and precision) of facial scanners using different technologies and the reliability of the texture maps produced.

Materials and Methods

A volunteer was scanned using nine scanners: Cloner, Artec MHT, Revopoint Mini, Revopoint POP 2, Vectra H2, EXAscan, Scaniverse app, iPhone 13 Pro MAX, and Nikon Z 7II Camera. Eighty scans were compared to a reference model (Vectra XT). Root mean square and standard deviation of the dimensional discrepancies were analyzed using ANOVA and Tukey post-hoc test. Acquisition times were assessed using Kruskal–Wallis test followed by Mann–Whitney U-test. Two expert assessments on texture map realism were averaged and compared to the reference using one sample t-test. The two experts' intraclass correlation coefficient (ICC) was calculated.

Results

Eight scanners fully captured the three-dimensional facial geometry. Overall trueness and precision of the facial scanners were significantly different (p < 0.001). Seven scanners showed clinically acceptable scanning accuracy of less than or equal to 2 mm. Cloner displayed the lowest deviation (0.61 ± 0.08 mm), fastest acquisition time (0.40 ± 0.00 s), and was highly reliable (<1 mm). The texture map of the Vectra H2 had the highest mean score of 8.50 and was the most realistic. The ICC between the experts was 0.78 indicating good interexaminer reliability.

Conclusions

The more economical facial scanners within each scanning technology (iPhone 13 Pro MAX, Cloner, Revopoint MINI, and Scaniverse app) showed clinically acceptable accuracy with realistic texture maps for facial scanning and therefore could be eligible substitutes for professional higher-cost scanners.