Trueness and scanning time of smartphone-based photogrammetry applications: an in vitro comparative study.

OBJECTIVE: To compare the global and regional trueness, native-scale usability, and full-face scanning time of three smartphone-based photogrammetry applications. METHODS: Thirty mannequin-head scans (n=10/application) were acquired with an iPhone 17 Pro using Polycam v.5.2.0, Qlone Dental v.1.0.14, and KIRI Engine v.4.1.1 under an identical 60-frame protocol. An ATOS Q scan served as the metrological reference. Global and facial-third RMS cloud-to-mesh deviations were calculated after scale-adjusted iterative closest point registration. Scanning time and the need for external rescaling were recorded (α=.05; Tukey pairwise tests). RESULTS: Global RMS trueness did not differ significantly among applications (P=.059; η²=.189). Qlone Dental showed lower RMS deviations than Polycam in the upper (0.45 ±0.17 mm versus 0.76 ±0.37 mm; P=.025) and lower facial thirds (0.68 ±0.07 mm versus 0.97 ±0.30 mm; P=.009). Only Qlone Dental in the upper third had a 95% confidence interval entirely below the 0.6-mm virtual-facebow benchmark. All Qlone Dental exports retained metric scale, whereas all Polycam and KIRI Engine exports required rescaling. Qlone Dental had the shortest scanning time (13.17 ±1.08 seconds; P<.001). CONCLUSIONS: Under standardized static conditions, none of the applications reliably achieved global RMS values below the 0.6-mm benchmark. The tested applications should not presently be considered standalone metric references for virtual facebow transfer; their more defensible role is in visualization, documentation, and communication workflows. CLINICAL SIGNIFICANCE: A controlled same-device comparison identified application-dependent trade-offs in regional reconstruction, native-scale usability, and capture speed.