Evaluation of the accuracy and usability of facial scans obtained with smartphones by different users.

Background: This study aims to compare 3-dimensional facial scan images obtained by patients and clinicians using a smartphone with the patients' actual facial dimensions.

Methods: With the patient's head in a stabilized position, patients were informed about the facial scanning process and instructed to perform a self-scan (group 1). The second scan was conducted by a clinician (group 2). Scanning times were recorded in groups 1 and 2. Finally, 17 direct anatomic measurements were taken from the patients (group 3). The acquired scan images were transferred to computer software, in which the anthropometric measurements were recalibrated and repeated virtually. The data obtained were statistically evaluated using the Shapiro-Wilk, t, Kruskal-Wallis, and Bonferroni tests.

Results: According to Bonferroni tests, statistically significant differences were observed in groups 1 and 3 between Sto-Ls and Sn-Me', Al(R)-Al(L), T(L)-Prn, Go(R)-Go(L), Sto-Me'; between Go(R)-Go(L) and Sto-Me', Ch(R)-Ch(L), Sto-Li; and between Sto-Me' and Sn-Sto, N-Sn groups. In groups 2 and 3, significant differences were noted between Sto-Ls and En(R)-En(L), Ex(R)-Ex(L), Sn-Me', Zyg(R)-Zyg(L), Sto-Me', Go(R)-Go(L); and between Sto-Li and Ex(R)-Ex(L), Sub-Me', Zyg(R)-Zyg(L), Sto-Me', Go(R)-Go(L) (P <0.05). In addition, in groups 1 and 3, measurements such as Sto-Me', Go(R)-Go(L), Me'-C, and T(L)-Prn showed an average difference of ≥1.00 mm between groups. In groups 2 and 3, Go(R)-Go(L), Sto-Me', Zyg(R)-Zyg(L), and Me'-C measurements showed an average difference of >1.00 mm between groups.

Conclusions: Facial scan images obtained with smartphones show an acceptable level of similarity to real measurements. Subjects without experience in facial scanning were able to perform facial scans at a high level of accuracy with the given instructions.