Influence of facial hair (clean-shaven, stubble, or beard) on the accuracy of 3-dimensional patient representation across four facial scanners: A clinical study.
TL;DR
Clinicians should consider the combined effect of facial hair style and the type of FS used during smiling facial scans, as it can influence the accuracy of VPR.
Abstract
STATEMENT OF PROBLEM Facial scans, when combined with intraoral scans, enable the creation of 3-dimensional virtual patient models that support comprehensive dentofacial analysis. However, the increasing prevalence of facial hair among patients may interfere with image acquisition and stitching processes. The extent to which facial hair affects the accuracy of virtual patient representation remains unclear.
Purpose
The purpose of this observational clinical study was to evaluate the influence of facial hair (clean-shaven, stubble, or beard) and the use of a dedicated beard-scanning mode on the accuracy of virtual patient representations (VPRs) obtained using professional and smartphone-based facial scanners (FSs). MATERIAL AND
Methods
Ten male participants were scanned using 3 professional (MetiSmile, Rayface, Morpheus) and 1 smartphone-based (Qlone) FSs. For each patient, facial scans were acquired under 2 conditions: with a full smile and with lip retraction (retracted smile). Prior to facial scanning, 4 fiducial landmarks were placed on predefined facial regions. The resulting facial datasets were subsequently aligned with intraoral scans to generate VPRs. A total of 20 linear distances between selected facial landmarks and tooth-based reference points were measured manually using digital calipers to establish reference values. The same measurements were then obtained digitally using a metrology software program. Statistical analysis was conducted using 2-way analysis of variance (ANOVA) for pairwise comparisons (α=.05).
Results
The effect of facial hair style on accuracy was significant only when VPRs were based on smiling facial scans and was influenced by the type of FS evaluated. VPRs derived from stubble facial hair using Metismile FS and from clean-shaven facial scans using Morpheus FS demonstrated significantly better trueness compared with the other evaluated professional and smartphone-based FSs. Facial hair style did not significantly affect accuracy, regardless of the FS used for VPRs based on retracted smile (P>.05). In contrast, facial hair style did not significantly affect accuracy when VPRs were based on retracted smile scans, regardless of the FS used (P>.05). For Metismile FS, activation of the dedicated beard scan mode improved trueness for full beard conditions but worsened trueness for stubble facial hair.
Conclusions
Clinicians should consider the combined effect of facial hair style and the type of FS used during smiling facial scans, as it can influence the accuracy of VPR.