Three-dimensional facial scanning vs. CBCT: a multi-dimensional pilot study

To subjectively assess perceived similarity, comfort, and convenience and objectively evaluate accuracy (trueness and precision) and clinical applicability (time and retake rate) of static and portable stereophotogrammetry (SPG), portable structured light (SL) scanners, front- and rear-camera smartphone apps. Facial scans of 10 patients were captured with static and portable SPG, portable SL scanners, and front- and rear-camera smartphone apps. Scans were registered and compared to soft tissue segmented from cone-beam computed tomography using surface-based analysis for accuracy assessment. Perceived similarity was rated immediately after scanning by both patients and operators using a three-point scale, with patients rating comfort. Similarity was further compared to two-dimensional photographs by 13 observers, and observer reliabilities were calculated. Scanning time, processing time, retake rate, and convenience were evaluated by the two operators who captured images. Both SPG and SL scanners demonstrated good trueness and precision, with portable SL showing higher precision values. All scanners and applications achieved clinically acceptable accuracy. Retake rates were lowest for static SPG (0%), followed by portable SPG (2%), portable SL (5%), and rear-camera app (7%) and highest for front-camera app (20%). For total time, front-camera app was the fastest, followed by static SPG, portable SL, portable SPG, and rear-camera app. Both patients and operators rated SPG and rear-camera app highest for similarity. All scanners achieved good similarity scores from observers, with SPG rated most favorably. SPG and both apps received good ratings for comfort and convenience, whereas other devices were rated fair. All scanners showed clinically acceptable accuracy and applicability, each with distinct strengths and limitations. SPG and SL performed best overall. Front-camera app was the fastest but with lower similarity perception, whereas rear-camera app showed the lowest accuracy but with high perceived similarity. Scanner selection should balance performance, perception, and specific clinical needs.