Purpose: To compare trueness and precision between conventional and digital facial measurements and to evaluate the accuracy of different superimposition techniques for facial scans.
Materials and methods: Twenty volunteers were recruited. Predetermined facial landmarks were marked with a black pen, and the interlandmark distances were measured manually with a conventional caliper and digitally with Geomagic software. Two consecutive facial scans were performed and superimposed using as best-fit reference the full face, the face without the eyes, and the bone-supported areas (eg, forehead and zygomatic areas) in order to assess root mean square (RMS) differences. Trueness and precision were evaluated and compared between the conventional and digital techniques. Mann-Whitney U and Kruskal-Wallis post hoc tests were used. The significance level was established at α = .05.
Results: Trueness between conventional and digital measurements was 1,151.75 ± 1,265.52 μm (3.04% ± 4.82%), and precision was 322.31 ± 300.54 μm (0.93% ± 1.10%). Global mean RMS values for each superimposition technique were 334.15 ± 172.07 for the full face, 339.57 ± 173.13 for the face without the eyes, and 385.65 ± 182.29 for the bone-supported areas, with the latter presenting statistically significant differences compared to the other two.
Conclusion: Although statistically significant differences were detected in facial measurements, they were below the clinically detectable threshold. Superimposition with the full face and the face without the eyes area presented smaller discrepancies than with the bone-supported areas, with higher discrepancies in the lower third of the face.