모바일 메뉴
Search
Search

KJO Korean Journal of Orthodontics

Open Access

pISSN 2234-7518
eISSN 2005-372X

퀵메뉴 버튼

Article

home All Articles View
Split Viewer

Original Article

Korean J Orthod 2025; 55(1): 37-47   https://doi.org/10.4041/kjod24.134

First Published Date October 23, 2024, Publication Date January 25, 2025

Copyright © The Korean Association of Orthodontists.

Are different photogrammetry applications on smartphones sufficiently reliable?

Gülden Karabiber , Hanife Nuray Yılmaz , Gamze Yıldırım

Department of Orthodontics, Marmara University, Istanbul, Türkiye

Correspondence to:Gülden Karabiber.
Assistant Professor, Department of Orthodontics, Marmara University, Başıbüyük Yolu No: 9/3 Başıbüyük, Istanbul 34854, Türkiye.
Tel +90-533-6261802 e-mail guldenkarabiber@hotmail.com

How to cite this article: Karabiber G, Yılmaz HN, Yıldırım G. Are different photogrammetry applications on smartphones sufficiently reliable? Korean J Orthod 2025;55(1):37-47. https://doi.org/10.4041/kjod24.134

Received: June 27, 2024; Revised: September 30, 2024; Accepted: October 18, 2024

Abstract

Objective: This study aimed to compare the accuracy of Qlone, Magiscan, and 3dMD with that of direct anthropometry (DA). Methods: The study involved 41 patients. Sixteen facial landmarks, including six individual and five paired points, were marked on each participant’s face. Subsequently, 18 linear measurements were assessed using a 3dMD device (multicamera photogrammetry), Qlone, Magiscan smartphone applications (single-camera photogrammetry), and DA. The Qlone and Magiscan images were calibrated using a reference point 10 mm from the nasion during DA to ensure a 1:1 correspondence. Results: Concerning the precision of the digital methods compared to DA, the mean intraclass correlation coefficient values of 3dMD, Qlone and Magiscan were 0.989, 0.980 and 0.982, respectively. Compared with DA, 3dMD achieved excellent trueness with the lowest average absolute differences in the measurements (highest value = 0.95 ± 0.62 mm). The highest values for Qlone and Magiscan were 1.51 ± 1.11 mm and 2.14 ± 1.69 mm, respectively. According to the number of parameters, the ranking of unreliable values (> 2 mm) was Magiscan (n = 46), Qlone (n = 35), and then, 3dMD (n = 4). Furthermore, reliability (less than 1 mm) was the highest for 3dMD (n = 517), followed by Magiscan (n = 457), and then, Qlone (n = 415). Conclusions: The 3dMD achieved excellent trueness with the lowest average absolute differences in the measurements. Based on statistical analysis, the trueness values of Magiscan and Qlone were close to that of 3dMD. To apply these smartphone applications clinically, more studies are necessary.

Keywords: Photography, 3-dimensional diagnosis, 3D scanner, Soft tissue

INTRODUCTION

Over the past few decades, facial esthetic analysis has gained significant importance in orthodontics. Orthodontists frequently rely on photographs to aid in esthetic diagnosis and treatment outcome evaluation.1,2 Traditional 2D photographs lack the accuracy needed to capture 3D facial details that are critical to comprehensive assessment. Hence, 3D photogrammetry is used.3,4 Historically, direct anthropometry (DA) was the gold standard for facial measurements owing to its affordability and reliability. However, DA is time consuming and operator dependent. Furthermore, it causes patient discomfort and data storage difficulties.5 To enhance measurement reliability and mitigate procedural errors, dentistry has shifted to 3D digital evaluation using facial scanning devices in recent years.6 3D imaging has many applications including pre-treatment evaluation of craniofacial deformities, cleft lip and palate, orthognathic surgery, syndromes, and facial asymmetries,7 as well as soft tissue and orthodontic treatment outcome evaluations.

3D digital imaging has been used in medicine since 1922.8 Various 3D facial scanning technologies, such as laser scanning, structured light scanning, stereophotogrammetry (multi-camera photogrammetry), photogrammetry (single-camera photogrammetry), and dual-structured light scanning, have been employed.9 The accuracy of laser scanning is hindered by lengthy scanning periods that may result in motion artifacts. All these systems are non-invasive, accurate and reproducible. Most systems require calibration (except for dual-structured light).10 The 3dMDface system (3dMD; 3dMD Inc., Atlanta, GA, USA) is widely used. It incorporates passive and active stereophotogrammetry in 3D surface imaging systems.11,12 Stereophotogrammetry offers several advantages. For example, the results are independent of light intensity and only one scan is required. 3dMD involves a multiple-camera system consisting of two modules, each with six cameras, and the shooting process does not require image stitching. The ability to capture an image in one scan prevents problems caused by patient movement.13 The disadvantages include high cost, large physical size, immobility, and its need for daily calibration.10,14

Most 3D photogrammetry systems, such as 3dMD and VECTRA M3 (3D Imaging System; Canfield Scientific, Parsippany, NJ, USA) are static devices. Currently, portable 3D imaging devices, such as VECTRA H1 and H2 (Canfield Scientific) comprise one digital single-lens reflex camera and a computer system. Typically, three consecutive photographs are taken from three angles, and then, merged to produce one 3D photo using computer software. Unlike static devices, portable instruments are relatively inexpensive and access is not location dependent. However, the reliability of portable devices is marginally lower than that of static devices.15,16

Photogrammetry is another widely used measurement technique. Although less expensive than stereophotogrammetry, photogrammetry requires several photographs and reverse engineering software for constructing a 3D image, while also being sensitive to light.10 More recently, cost-effective 3D scanning systems have emerged, including the dual-structured light system Bellus3D FaceMaker (Bellus 3D Inc., Campbell, CA, USA), which can be installed as an application on smartphones or tablets. This application facilitates 3D facial photography using an external camera integrated into a smartphone or tablet. Importantly, the accuracy of Bellus3D is similar to that of 3dMD, and the cost difference is almost negligible.14,17

Other smartphone 3D scanning applications include Qlone, Magiscan, EM3D, Heges, ScandyPro, Polycam, Widar, Kiri Engine, and Trnio.18 Of these, Qlone (version 5.6.0; EyeCue Vision Technologies, LTD-AR Technologies, Yokne’am, Israel) and Magiscan (version 1.5.12; AR Generation, AR and LiDAR Technologies, Warsaw, Poland) are the most user-friendly for clinical applications. Hence, we included these two applications into our study. To the best of our knowledge, the accuracy of these new applications has not been previously reported.

Magiscan and Qlone are 3D scanning applications that use single-camera photogrammetry and capture object distances using laser pulses, similar to a radar. An infrared transmitter on the mobile light detection and ranging (LiDAR) camera projects dots on the scanning surface,19 and thus, direct daylight impacts the scanning quality of LiDAR cameras.20 In contrast, Magiscan needs LiDAR sensors (compatible with iPhone 12 Pro and above), whereas Qlone employs photogrammetry without the need for LiDAR. Both applications use photogrammetry to convert a series of pictures taken on an iPhone or iPad into USDZ files that can be used to construct a 3D image or viewed in augmented reality. USDZ is Apple’s file format for augmented reality on the iOS platform. Meanwhile, the extended image acquisition time associated with these applications, compared with 3dMD, may lead to complications arising from patient movement. Assessment of currently available 3D photography applications is necessary to facilitate the integration of these novel methods into dentistry and orthodontics.

Deutsch et al.21 defined the anatomical landmarks for anthropometric measurements. In studies examining reliability, careful selection of landmarks is crucial, especially those affected by facial expressions (e.g., perioral points), with unclear visualization (e.g., tragus), or those that are difficult to detect anatomically (e.g., gonion). In the literature, landmarks located on the ears, such as tragions, are difficult to place precisely because of potential shadows and hair obstructions. The 3dMDface and Bellus 3D systems produce surface deviations around the ears; thus, the accuracy of ear landmarks is insufficient.16,22 If those areas must be measured, the 3dMD system may increase the precision of the landmarks on the ears.

The present study aimed to compare the accuracy of Qlone, Magiscan, and 3dMD, and DA. We hypothesized that photogrammetry systems on smartphones are reliable and cost-effective alternatives to stereophotogrammetry (3dMD).

MATERIALS AND METHODS

This single-center, single-blinded study was approved by the Ethical Committee of Marmara University Dentistry Faculty Clinical Researches Ethical Committee (approval date and protocol number: 26.01.2023 and 2023/121) and was conducted according to the Declaration of Helsinki (World Medical Association, 2013). Sample size calculations were performed using the G*Power program (version 3.1.9.2; Heinrich-Heine-University, Düsseldorf, Germany) with a statistical power of 90%, an alpha error of 0.05, and an effect size of 0.3. The sample population consisted of 41 patients, with a mean age of 24.74 ± 3.16 years. Participants were informed verbally and in writing before the study. Informed consent was obtained from all participants. The inclusion criteria were adults, students or staff of the School of Dentistry, aged 18–35 years, stable occlusion, willingness to participate, and signed informed consent. Individuals with dentofacial deformities, facial trauma, surgical scars, skin diseases, mustaches, beards, or acne covering facial landmarks were excluded.

Data collection

Initially, DA measurements were performed using the standard anthropometric landmarks defined by Deutsch et al.21 Of the 32 landmarks, 16 that were the least affected by gestures and movements, were selected. The selected landmarks (6 single and 5 paired) were as follows: reference point (ref), zygonion (zygo right [R]-left [L]), trichion (tri), nasion (n), subnasale (sn), Pronasale (prn), endocanthion (endo R-L), eyebrow (eb R-L), alare (ala R-L), pogonion (pog). These were initially marked on each individual’s face using an acetate pen with a 0.5 mm tip (Figure 1). Eighteen linear measurements based on the landmarks, including transverse, vertical, sagittal oblique, and vertical oblique measurements were evaluated (Table 1).

Figure 1. The landmarks used in the present study.
See Table 1 for definitions of each landmark or measurement.

Table 1 . Description of anthropometric landmarks and definition of linear distances

DirectionAbbreviationDescription
TransversalendoR-endoLDistance between right and left endocanthions (inner corner of eye)
alaR-alaLDistance between right and left alares (most prominent point of alare)
ebR-ebLDistance between right and left initial point of eyebrows
zygoR-zygoLDistance between right and left zygonions (most prominent point in zygomatic region)
Sagittal-ObliquezygoL-alaLDistance between left zygonion and alare
zygoR-alaRDistance between right zygonion and alare
alaL-prnDistance between left alare and pronasale
alaR-prnDistance between right alare and pronasale
zygoL-prnDistance between left zygonion and pronasale
zygoR-prnDistance between right zygonion and pronasale
Verticaltri-nDistance between trichion and nasion
n-snDistance between nasion and subnasale
sn-pogDistance between subnasale and pogonion
n-prnDistance between nasion and pronasale
Vertical-ObliquezygoL-nDistance between left zygonion and nasion
zygoR-nDistance between right zygonion and nasion
zygoL-pogDistance between left zygonion and pogonion
zygoR-pogDistance between right zygonion and pogonion

endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left.



The patients were seated on a height-adjustable chair and instructed to look at a mirror fixed in front of them to establish a natural head position. Images acquired from 3dMD were automatically saved in a tricorder surface binary (.tsb) format. Additionally, single-camera photogrammetry was performed using the smartphone applications, namely, Qlone and Magiscan. To construct the 3D images, Qlone and Magiscan captured 40 and 45 serial photographs per scan, respectively. 3D scans were obtained with the patient in a stable head position, natural expression, and maximum intercuspation while the camera moved around the head.17 All the scans were performed using a 48-megapixel True depth camera, which uses LiDAR, with an A16 Bionic chip (iPhone 14 Pro, iOS 16.1.1, Apple Store, Cupertino, CA, USA). After the scanning procedure, the 3D files were saved in the .obj format (Figure 2).

Figure 2. A, 3dMD image. B, Magiscan image. C, Qlone image.

Due to surface deviations around the ears, precisely placing landmarks in that region was difficult, resulting in insufficient accuracy. Consequently, the ear regions were removed from all images. To provide a visual representation, a sample color map was generated by superimposing three digital samples (Figure 3).

Figure 3. A 3D color map of the superimposed samples for a patient.

Data measurement

DA measurements were performed using a digital caliper (0–150 mm, Digital Caliper 1112-150, INSIZE Co., Ltd., Suzhou New District, China) with an accuracy of 0.03 mm. For the digital measurements, .obj files were imported into the free-source MeshLab software (ISTI [Italian National Research Council], Rome, Italy). A reference point was marked 10 mm above the nasion to calibrate the scanned photographs (Figure 1). The mean differences between the Qlone and Magiscan images and the DA were scaled using coefficients calculated based on a 10 mm scale, resulting in a 1:1 correspondence. Of note, the magnification factor for Magiscan was approximately 200, and approximately 1.1 to 1.3 times for Qlone. Therefore, a minor error may be amplified during the initial measurement owing to the multiplication of the coefficient.

All measurements were repeated on the same day by the same operator to calculate intra-examiner repeatability. Accuracy is the agreement between a measurement and the “true” value of a parameter, as provided by the DA and measured using a caliper. The average absolute distance between the DA and digital inter-landmark distances was assessed to evaluate accuracy. Additionally, the mean values of the parameters for each measurement method were compared to assess their reliability.

Statistical analysis

IBM SPSS Statistics (version 22.0; IBM Corp., Armonk, NY, USA) was used for all the statistical analyses. The conformity of the parameters to the normal distribution was assessed using the Shapiro–Wilk test. To evaluate the study data, the intraclass correlation coefficient (ICC) was calculated to determine the reliability of the digital methods compared to DA. The Friedman test was used to evaluate the amount of deviation of the other scanners compared to the direct anthropometric measurements. Bonferroni correction was performed for pairwise comparison. Intra-examiner reliability was assessed using the ICC. Statistical significance was set at P < 0.05.

RESULTS

The intra-examiner reliability of DA, 3dMD, Qlone, and Magiscan ranged from 0.990 to 1.000, 0.997 to 1.000, 0.991 to 1.000, and 0.981 to 1.000, respectively, indicating a high level of agreement. The mean values of each parameter for each measurement method are presented in Table 2. Concerning the precision of the digital methods compared with DA, the mean ICC values of the 3dMD, Qlone, and Magiscan were 0.989, 0.980, and 0.982, respectively. The lowest ICC was 0.972 for 3dMD, 0.950, and 0.949 for Qlone (Table 2).

Table 2 . Descriptive statistics of the variables for each method

DA3dMDQloneMagiscanDA-3dMDDA-QloneDA-Magiscan
Mean ± SDMean ± SDMean ± SDMean ± SDICC (95% CI)ICC (95% CI)ICC (95% CI)
Transversal
endoR-endoL (mm)28.83 ± 2.4828.97 ± 2.5628.92 ± 2.5728.93 ± 2.380.976 (0.956–0.987)0.960 (0.925–0.979)0.984 (0.970–0.991)
alaR-alaL (mm)32.70 ± 2.3932.91 ± 2.5133.33 ± 2.6432.89 ± 2.450.972 (0.948–0.985)0.949 (0.904–0.973)0.978 (0.959–0.988)
ebR-ebL (mm)20.47 ± 4.3520.23 ± 4.2420.27 ± 4.4020.28 ± 4.240.995 (0.991–0.977)0.995 (0.991–0.998)0.994 (0.989–0.997)
zygoR-zygoL (mm)121.70 ± 7.80121.42 ± 7.79121.3 ± 7.97120.01 ± 7.780.996 (0.992–0.998)0.986 (0.974–0.993)0.980 (0.963–0.990)
Sagittal-Oblique
zygoR-alaR (mm)55.16 ± 6.1954.60 ± 6.1154.68 ± 6.1554.50 ± 6.130.997 (0.993–0.998)0.992 (0.985–0.996)0.992 (0.985–0.996)
zygoL-alaL (mm)59.35 ± 4.8058.92 ± 4.9859.08 ± 4.9958.74 ± 5.050.993 (0.987–0.996)0.998 (0.997–0.994)0.989 (0.979–0.994)
alaR-prn (mm)27.26 ± 1.9226.40 ± 1.7826.95 ± 1.9526.80 ± 1.890.974 (0.951–0.986)0.958 (0.921–0.977)0.959 (0.922–0.978)
alaL-prn (mm)27.03 ± 1.9026.39 ± 1.8826.74 ± 2.0626.69 ± 1.960.972 (0.948–0.985)0.962 (0.929–0.980)0.950 (0.906–0.973)
zygoR-prn (mm)79.53 ± 7.0978.61 ± 7.0379.25 ± 7.1978.47 ± 6.870.997 (0.995–0.998)0.991 (0.983–0.995)0.993 (0.986–0.996)
zygoL-prn (mm)83.75 ± 5.7082.91 ± 5.7383.27 ± 5.9382.66 ± 5.820.996 (0.992–0.998)0.982 (0.966–0.990)0.983 (0.968–0.991)
Vertical
tri-n (mm)63.15 ± 7.5063.40 ± 7.4663.28 ± 7.4062.86 ± 7.550.996 (0.993–0.998)0.993 (0.987–0.996)0.992 (0.986–0.996)
n-sn (mm)51.37 ± 3.6351.63 ± 3.5351.63 ± 3.6051.27 ± 3.460.987 (0.975–0.993)0.977 (0.957–0.998)0.981 (0.965–0.990)
sn-pog (mm)50.68 ± 4.2551.17 ± 4.3851.23 ± 4.3250.92 ± 4.260.989 (0.979–0.994)0.980 (0.963–0.989)0.984 (0.969–0.991)
n-prn (mm)43.10 ± 4.3943.19 ± 4.2743.37 ± 4.5042.90 ± 4.300.996 (0.992–0.998)0.989 (0.980–0.994)0.992 (0.985–0.996)
Vertical-Oblique
zygoR-n (mm)73.76 ± 5.4673.26 ± 5.4773.58 ± 5.4573.07 ± 5.450.993 (0.986–0.996)0.987 (0.976–0.993)0.986 (0.973–0.992)
zygoL-n (mm)75.05 ± 4.5774.57 ± 4.5675.00 ± 4.6274.34 ± 4.630.994 (0.998–0.997)0.981 (0.965–0.990)0.978 (0.959–0.988)
zygoR-pog (mm)96.98 ± 6.4496.86 ± 6.4097.09 ± 6.8396.19 ± 6.600.996 (0.992–0.998)0.989 (0.980–0.994)0.989 (0.980–0.994)
zygoL-pog (mm)100.56 ± 6.13100.95 ± 6.34100.82 ± 6.34100.02 ± 6.100.995 (0.991–0.997)0.981 (0.964–0.990)0.985 (0.971–0.992)

ICC, intraclass correlation coefficient; DA, direct anthropometry; SD, standard deviation; CI, confidence interval; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left.



Table 3 presents the average absolute differences in the linear distances between the DA and digital methods. Compared with DA (the reference standard for the present study), 3dMD achieved high accuracy and the lowest average absolute difference, with the highest value being 0.95 ± 0.62 mm. Conversely, Qlone and Magiscan showed greater differences relative to DA. Moreover, Qlone showed significantly higher average absolute differences in the parameters of endoR-endoL (0.80 ± 0.59 mm), alaR-alaL (1.05 ± 0.72 mm), n-prn (0.77 ± 0.57 mm), zygoL-n (1.08 ± 0.60 mm), and zygoL-pog (1.42 ± 0.96 mm), whereas the highest significant value in Magiscan was observed for zygoR-zygoL (2.14 ± 1.69 mm) (P < 0.05).

Table 3 . Average absolute differences between DA and other scanning methods and pairwise comparisons of the average absolute differences between the methods

ParameterDA-3dMDDA-QloneDA-MagiscanPa3dMD-Qlone
(Pb)
3dMD-Magiscan
(Pb)
Qlone-Magiscan
(Pb)
Mean ± SDMean ± SDMean ± SD
Transversal
endoR-endoL (mm)0.61 ± 0.480.80 ± 0.590.47 ± 0.400.049*1.0000.3660.045*
alaR-alaL (mm)0.67 ± 0.481.05 ± 0.720.44 ± 0.570.000***0.0610.0610.000***
ebR-ebL (mm)0.49 ± 0.410.43 ± 0.450.51 ± 0.430.438---
zygoR-zygoL (mm)0.89 ± 0.541.51 ± 1.112.14 ± 1.690.001***0.0820.001***0.502
Sagittal-Oblique
zygoR-alaR (mm)0.79 ± 0.481.00 ± 0.671.05 ± 0.700.109---
zygoL-alaL (mm)0.75 ± 0.510.89 ± 0.650.98 ± 0.700.181---
alaR-prn (mm)0.89 ± 0.540.63 ± 0.540.74 ± 0.490.071---
alaL-prn (mm)0.67 ± 0.580.65 ± 0.470.76 ± 0.490.426---
zygoR-prn (mm)0.95 ± 0.621.15 ± 0.771.32 ± 0.880.787---
zygoL-prn (mm)0.93 ± 0.611.20 ± 1.081.45 ± 1.130.067---
Vertical
tri-n (mm)0.77 ± 0.520.94 ± 0.771.06 ± 0.810.249---
n-sn (mm)0.70 ± 0.500.93 ± 0.610.81 ± 0.510.081---
sn-pog (mm)0.92 ± 0.441.05 ± 0.780.93 ± 0.590.843---
n-prn (mm)0.44 ± 0.370.77 ± 0.570.65 ± 0.470.028*0.024*0.9610.293
Vertical-Oblique
zygoR-n (mm)0.88 ± 0.591.08 ± 0.591.06 ± 0.990.232---
zygoL-n (mm)0.70 ± 0.521.08 ± 0.601.16 ± 0.980.021*0.017*0.3660.673
zygoR-pog (mm)0.67 ± 0.491.08 ± 0.801.17 ± 1.030.157---
zygoL-pog (mm)0.78 ± 0.541.42 ± 0.961.13 ± 1.120.002**0.002**0.9610.039*

DA, direct anthropometry; SD, standard deviation; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left.

aFriedman test, bBonferroni correction for multiple tests.

*P < 0.05, **P < 0.01, ***P < 0.001.



The mean absolute differences of the linear measurements for 3dMD, Qlone, and Magiscan were 0.75 ± 0.53 mm, 0.98 ± 0.77 mm, and 0.99 ± 0.92 mm, respectively. Figure 4 shows the reliability of the methods according to the average absolute differences for all the parameters (738 in total). Notably, 3dMD demonstrated four unreliable measurements, while Qlone and Mgiscan showed 35 and 46, respectively.

Figure 4. Reliability differences among 3dMD, Magiscan, and Qlone.

DISCUSSION

This study compared the accuracy of the Qlone, Magiscan, and 3dMD systems to each other and to the DA method. During the preliminary stages of the research design, multiple smartphone applications were evaluated, among which EM3D, Heges, and ScandyPro generated 3D images by employing a video-recording approach analogous to that of Bellus 3D. EM3D and Heges produced images using a lower-megapixel front camera, while ScandyPro used the rear camera.18 All of these applications demonstrated challenges regarding accurate image capture and tracking loss. Considering that 3D scanning technologies depend on photogrammetry, the 3D images created in Magiscan and Qlone by combining serial photos taken using a smartphone’s high-megapixel rear camera produced better image quality for the construction of 3D photographs without tracking loss. Moreover, the selected applications were the most cost efficient.

Based on empirical observations, optimal 3D image acquisition with those applications required a well-illuminated environment with direct daylight exposure on the participant’s face.18 This divergence from 3dMD’s performance characteristics may be attributed to the light sensitivity of the applications.10

In the present study, angular measurements were excluded because of the inherent challenges associated with DA stemming from operator experience and soft tissue resilience. Additionally, changes in facial expressions may affect the positions of landmarks in the perioral region, and it is difficult to precisely locate the gonion without direct palpation. Therefore, these landmarks were excluded from the study.

For the ICC, a value greater than 0.900 indicated that the measurements were consistent and reproducible. Most previous studies reported ICC values higher than 0.900 for 3dMD,5,17,23,24 Dental Pro,25 Face Hunter,25 and Bellus 3D.17,23,24 Comparing the DA and digital methods, all measurements showed high agreement, with ICC values greater than 0.900. Among the evaluated methods, 3dMD presented the highest ICC values for most parameters, followed by Magiscan, and then Qlone. The lowest ICC value was 0.972 for 3dMD, 0.950 for Magiscan, and 0.949 for Qlone. These results indicate that each method was highly reliable and precise compared with DA.

Liu et al.17 reported a mean absolute difference of 0.36 ± 0.20 mm for 3dMD. In the present study, the clinical acceptance has been set at 1 mm. The accuracy of 3dMD was found to be clinically acceptable5 and showed a mean absolute difference of 0.75 ± 0.53 mm compared with DA. Additionally, Qlone and Magiscan achieved mean absolute differences of 0.98 ± 0.77 mm and 0.99 ± 0.92 mm, respectively, which are considered generally reliable and within acceptable limits (except for the zygonion measurements). This may be explained by the lateral positioning of the zygo to the midline, which is located on the curvature of the face, and the enlargement of that area with these techniques. Considering the relatively low accuracy associated with the zygonion landmark for these two methods, the exocanthion may be appropriate in future oblique measurement studies. However, the risk of squinting and displacing these points should not be ignored, especially in methods in which image acquisition takes a long time. In a previous study of the Dental Pro and Face Hunter applications, significant changes of the zygo R-L measurement was reported.25 Although similar differences were not found to be significant in the present study, differences over 1 mm were observed in the measurements including zygo landmarks, in the Magiscan and Qlone 3D images.

Additionally, the measurement that included the sn point showed a 0.70–1.05 mm difference using the three digital imaging techniques compared with DA. This is because the sn is difficult to locate due to the curvature and positional variations of the lip (including lip posture and morphology) and the nose.26 In 3D imaging, attention should be paid to parameters involving the sn. To summarize, larger differences were observed when acquiring areas with more curvature, especially when using Qlone and Magiscan. These results are supported by Hong et al.,27 who reported the greatest errors in areas with curvatures and larger variations, especially in longer measurements. Additional improvements, especially in smartphone applications, may be necessary to accurately capture the areas of curvature.

Dindaroğlu et al.5 reported a deviation ranging from −0.21 to 0.15 mm between DA and 3dMD. However, the deviation values were not average absolute differences, such as those calculated in the present study.5 This may be the reason why the results reported by Dindaroğlu et al.5 were lower than the results presented herein. Dindaroğlu et al.5 found a significant change in the endo R-L distance (ranging between −1.16 and 0.74 mm) when DA and 3dMD were compared. In the present study, this parameter was significant when the average absolute differences of Qlone (0.80 mm) and Magiscan (0.47 mm) compared to DA were considered, showing higher differences in Qlone.

Raffone et al.28 found the difference between the ala to be 0.788 mm for Bellus 3D. In this study, the ala R-L measured 0.67 mm for 3dMD, 1.05 mm for Qlone, and 0.44 mm for Magiscan. Additionally, the differences in the ala-prn distance between Bellus 3D and DA was 0.524 mm on the right side and 1.177 mm on the left side.28 The differences for those parameters in the present study were 0.89 mm for 3dMD, 0.63 mm for Qlone, and 0.74 mm for Magiscan on the right side; 0.67 mm for 3dMD, 0.65 mm for Qlone, and 0.76 mm for Magiscan on the left side. Moreover, the alar wings moved differently, depending on the breathing pattern during image acquisition. For instance, the alar wings of a nasal breather tended to expand during breathing, whereas an oral breather presents with narrowing or inactivity of the alar wings. This may help explain the large ala measurement differences in studies.

The average absolute difference in the n-prn between DA and 3dMD was found to be 0.44 mm, which was consistent with the results of previous reports (0.3–0.4 mm).17 Furthermore, Qlone and Magiscan exhibited a deviation of 0.77 mm and 0.65 mm, respectively. Raffone et al.28 compared the slider and free techniques. With the slider technique, the patient’s head is motionless and the scanner moves around the head. With the free technique, head is rotated, while the scanner remains stable. They reported that the slider technique was more suitable for clinical use due to fewer motion artifacts, thus, resulting in a scan free from distortions.28 In the present study, the images were obtained by employing the sliding technique.

The ICCs for the agreement between Bellus 3D and DA mostly fell below the acceptable limit of 0.80.24 In contrast, all the methods were found to be highly precise in the present study, presenting high ICC values compared with DA. According to the results, Magiscan and Qlone exhibited fewer differences than Bellus 3D (dual-structure light). However, studies that include the Bellus 3D and smartphone applications may provide more definitive and reliable results.

To clarify the results, a reliability table was constructed by employing a 2 mm reliability threshold. This categorization, referring to established studies, classified the measurements into four intervals.29 Although some studies considered a 2 mm discrepancy as clinically acceptable,28 others proposed a threshold of 1 mm.5,17 However, developing technology should promise to reduce this limit. Thus, we adopted a 1 mm deviation as the criterion for clinical acceptability.

This present study included 738 measurements. Of these, 76.28% and 81.43% of the Magiscan and Qlone measurements provided highly reliable outcomes, with 6.23% and 4.74% for Magiscan and Qlone, respectively. Comparable analyses in the literature found 84% of the measurements to be highly reliable and reliable, 5% to be unreliable using Dental Pro, 87% to be highly reliable and reliable, and 1% to be unreliable using Face Hunter.25

The post-capture image-processing time with photogrammetry in Magiscan (5–10 minutes) was notably longer than that in Qlone (1–2 minutes), which is a disadvantage for clinical application. Both the Magiscan and Qlone require environments with sufficient daylight. Finally, the possibility of deformation in pointed and curved areas of the face, such as the zygo and alar wings, should be considered. To date, available studies are limited and some have performed their analyses using a mannequin head.10,17,25,27,28 Since the mannequin head is not affected by movement, methods must be tested in humans to accurately determine its clinical applicability. After all, measurements must ultimately be performed on individuals in the clinical setting.

One limitation of the present study is the absence of inter-examiner reliability. The operator’s smartphone handling skills may significantly impact the results. In the future, more studies on smartphone application updates and the different scanning techniques are needed, so that more people can create consistent 3D images using the same program.

CONCLUSIONS

Qlone presented significantly higher values for the parameters n-prn, zygoL-n, and zygoL-pog, while the highest significant value for Magiscan was observed for ZygoR-zygoL compared with 3dMD.

The 3dMD achieved high accuracy and exhibited the lowest average absolute differences. Magiscan and QLONE are cost-effective alternatives to 3dMD. Hence, their integration into clinical routines may be recommended without ignoring the fact that Magiscan and Qlone have 10 times the number of unreliable measurements compared with 3dMD.

The absence of supplementary cameras and their economical pricing render Qlone and Magiscan advantageous compared to alternative systems. Further studies are required to validate clinical reliability.

The post-capture image processing time associated with photogrammetry using Magiscan was notably longer than that using Qlone, which is a disadvantage of 3dMD.

AUTHOR CONTRIBUTIONS

Conceptualization: GK, HNY. Data curation: GY. Formal analysis: GY. Investigation: GK. Methodology: GK, HNY. Project administration: GK, HNY. Resources: HNY. Software: GY. Supervision: GK, HNY. Validation: HNY, GY. Visualization: GK, GY. Writing–original draft: GK. Writing–review & editing: HNY.

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

FUNDING

None to declare.

References

  1. Nair AG, Santhanam A. Clinical photography for periorbital and facial aesthetic practice. J Cutan Aesthet Surg 2016;9:115-21. https://doi.org/10.4103/0974-2077.184047
    Pubmed KoreaMed CrossRef
  2. Stebel A, Desmedt D, Bronkhorst E, Kuijpers MA, Fudalej PS. Rating nasolabial appearance on three-dimensional images in cleft lip and palate: a comparison with standard photographs. Eur J Orthod 2016;38:197-201. https://doi.org/10.1093/ejo/cjv024
    Pubmed KoreaMed CrossRef
  3. Jyothikiran H, Shanthara JR, Subbiah P, Thomas M. Craniofacial imaging in orthodontics--past present and future. Int J Orthod Milwaukee 2014;25:21-6. https://pubmed.ncbi.nlm.nih.gov/24812737/
  4. Da Silveira AC, Daw JL Jr, Kusnoto B, Evans C, Cohen M. Craniofacial applications of three-dimensional laser surface scanning. J Craniofac Surg 2003;14:449-56. https://doi.org/10.1097/00001665-200307000-00009
    Pubmed CrossRef
  5. Dindaroğlu F, Kutlu P, Duran GS, Görgülü S, Aslan E. Accuracy and reliability of 3D stereophotogrammetry: a comparison to direct anthropometry and 2D photogrammetry. Angle Orthod 2016;86:487-94. https://doi.org/10.2319/041415-244.1
    Pubmed KoreaMed CrossRef
  6. Li Q, Bi M, Yang K, Liu W. The creation of a virtual dental patient with dynamic occlusion and its application in esthetic dentistry. J Prosthet Dent 2021;126:14-8. https://doi.org/10.1016/j.prosdent.2020.08.026
    Pubmed CrossRef
  7. Erten O, Yılmaz BN. Three-dimensional imaging in orthodontics. Turk J Orthod 2018;31:86-94. https://doi.org/10.5152/TurkJOrthod.2018.17041
    Pubmed KoreaMed CrossRef
  8. Mannsbach M. Die stereophotogrammetrie als hilfsmittel in der orthodontie. Dtsch Zahnärzt Orthop 1922;14:105. https://scholar.google.co.kr/scholar?hl=ko&as_sdt=0%2C5&q=%22Die+stereophotogrammetrie+als+hilfsmittel+in+der+orthodontie%22&btnG=
  9. Magnani M, Douglass M. Photogrammetry and stereophotogrammetry. In: López Varela SL, ed. The encyclopedia of archaeological sciences. Chichester: Wiley Blackwell; 2019. p. 1-4. https://doi.org/10.1002/9781119188230.saseas0451
    CrossRef
  10. Piedra-Cascón W, Meyer MJ, Methani MM, Revilla-León M. Accuracy (trueness and precision) of a dual-structured light facial scanner and interexaminer reliability. J Prosthet Dent 2020;124:567-74. https://doi.org/10.1016/j.prosdent.2019.10.010
    Pubmed CrossRef
  11. Yüksel Coşkun E, Esenlik E. A prospective study comparing adolescent and post-adolescent periods regarding effects of activator appliance in patients with class II mandibular retrognathia by using 3dMDface analysis and cephalometry. Med Sci Monit 2020;26:e921401. https://doi.org/10.12659/msm.921401
    Pubmed KoreaMed CrossRef
  12. White JD, Ortega-Castrillon A, Virgo C, Indencleef K, Hoskens H, Shriver MD, et al. Sources of variation in the 3dMDface and Vectra H1 3D facial imaging systems. Sci Rep 2020;10:4443. https://doi.org/10.1038/s41598-020-61333-3
    Pubmed KoreaMed CrossRef
  13. Zhao YJ, Xiong YX, Wang Y. Three-dimensional accuracy of facial scan for facial deformities in clinics: a new evaluation method for facial scanner accuracy. PLoS One 2017;12:e0169402. https://doi.org/10.1371/journal.pone.0169402
    Pubmed KoreaMed CrossRef
  14. Aynechi N, Larson BE, Leon-Salazar V, Beiraghi S. Accuracy and precision of a 3D anthropometric facial analysis with and without landmark labeling before image acquisition. Angle Orthod 2011;81:245-52. https://doi.org/10.2319/041810-210.1
    Pubmed KoreaMed CrossRef
  15. Fan W, Guo Y, Hou X, Liu J, Li S, Ju S, et al. Validation of the portable next-generation VECTRA H2 3D imaging system for periocular anthropometry. Front Med (Lausanne) 2022;9:833487. https://doi.org/10.3389/fmed.2022.833487
    Pubmed KoreaMed CrossRef
  16. Loy RCH, Liew MKM, Yong CW, Wong RCW. Validation of low-cost mobile phone applications and comparison with professional imaging systems for three-dimensional facial imaging: a pilot study. J Dent 2023;137:104676. https://doi.org/10.1016/j.jdent.2023.104676
    Pubmed CrossRef
  17. Liu J, Zhang C, Cai R, Yao Y, Zhao Z, Liao W. Accuracy of 3-dimensional stereophotogrammetry: comparison of the 3dMD and Bellus3D facial scanning systems with one another and with direct anthropometry. Am J Orthod Dentofacial Orthop 2021;160:862-71. https://doi.org/10.1016/j.ajodo.2021.04.020
    Pubmed CrossRef
  18. Van Lint L, Christiaens L, Stroo V, Bila M, Willaert R, Sun Y, et al. Accuracy comparison of 3D face scans obtained by portable stereophotogrammetry and smartphone applications. J Med Biol Eng 2023;43:550-60. https://doi.org/10.1007/s40846-023-00817-9
    CrossRef
  19. Harrap R, Daniel S. Mobile lidar mapping: building the next generation of outdoor environment models for AR [Internet]. IEEE International Symposium on Mixed and Augmented Reality ISMAR: Orlando, FL; 2009 [cited 2009 Oct 19]. Available from: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=7406fce8f7fc17a083eb533aafae2b91055d3961
  20. Akan B, Akan E, Şahan AO, Kalak M. Evaluation of 3D face-scan images obtained by stereophotogrammetry and smartphone camera. Int Orthod 2021;19:669-78. https://doi.org/10.1016/j.ortho.2021.08.007
    Pubmed CrossRef
  21. Deutsch CK, Shell AR, Francis RW, Bird BD. The farkas system of craniofacial anthropometry: methodology and normative databases. In: Preedy VR, ed. Handbook of anthropometry: physical measures of human form in health and disease. New York: Springer New York; 2012. p. 561-73. https://doi.org/10.1007/978-1-4419-1788-1_29
    CrossRef
  22. Aldridge K, Boyadjiev SA, Capone GT, DeLeon VB, Richtsmeier JT. Precision and error of three-dimensional phenotypic measures acquired from 3dMD photogrammetric images. Am J Med Genet A 2005;138a:247-53. https://doi.org/10.1002/ajmg.a.30959
    Pubmed KoreaMed CrossRef
  23. Weinberg SM, Naidoo S, Govier DP, Martin RA, Kane AA, Marazita ML. Anthropometric precision and accuracy of digital three-dimensional photogrammetry: comparing the Genex and 3dMD imaging systems with one another and with direct anthropometry. J Craniofac Surg 2006;17:477-83. https://doi.org/10.1097/00001665-200605000-00015
    Pubmed CrossRef
  24. Staller S, Anigbo J, Stewart K, Dutra V, Turkkahraman H. Precision and accuracy assessment of single and multicamera three-dimensional photogrammetry compared with direct anthropometry. Angle Orthod 2022;92:635-41. https://doi.org/10.2319/101321-770.1
    Pubmed KoreaMed CrossRef
  25. Pellitteri F, Brucculeri L, Spedicato GA, Siciliani G, Lombardo L. Comparison of the accuracy of digital face scans obtained by two different scanners. Angle Orthod 2021;91:641-9. https://doi.org/10.2319/092720-823.1
    Pubmed KoreaMed CrossRef
  26. Masoud MI, Bansal N, Castillo JC, Manosudprasit A, Allareddy V, Haghi A, et al. 3D dentofacial photogrammetry reference values: a novel approach to orthodontic diagnosis. Eur J Orthod 2017;39:215-25. https://doi.org/10.1093/ejo/cjw055
    Pubmed CrossRef
  27. Hong C, Choi K, Kachroo Y, Kwon T, Nguyen A, McComb R, et al. Evaluation of the 3dMDface system as a tool for soft tissue analysis. Orthod Craniofac Res 2017;20 Suppl 1:119-24. https://doi.org/10.1111/ocr.12178
    Pubmed KoreaMed CrossRef
  28. Raffone C, Gianfreda F, Bollero P, Pompeo MG, Miele G, Canullo L. Chairside virtual patient protocol. Part 1: free vs guided face scan protocol. J Dent 2022;116:103881. https://doi.org/10.1016/j.jdent.2021.103881
    Pubmed CrossRef
  29. Aung SC, Ngim RC, Lee ST. Evaluation of the laser scanner as a surface measuring tool and its accuracy compared with direct facial anthropometric measurements. Br J Plast Surg 1995;48:551-8. https://doi.org/10.1016/0007-1226(95)90043-8
    Pubmed CrossRef

Article

Original Article

Korean J Orthod 2025; 55(1): 37-47   https://doi.org/10.4041/kjod24.134

First Published Date October 23, 2024, Publication Date January 25, 2025

Copyright © The Korean Association of Orthodontists.

Are different photogrammetry applications on smartphones sufficiently reliable?

Gülden Karabiber , Hanife Nuray Yılmaz , Gamze Yıldırım

Department of Orthodontics, Marmara University, Istanbul, Türkiye

Correspondence to:Gülden Karabiber.
Assistant Professor, Department of Orthodontics, Marmara University, Başıbüyük Yolu No: 9/3 Başıbüyük, Istanbul 34854, Türkiye.
Tel +90-533-6261802 e-mail guldenkarabiber@hotmail.com

How to cite this article: Karabiber G, Yılmaz HN, Yıldırım G. Are different photogrammetry applications on smartphones sufficiently reliable? Korean J Orthod 2025;55(1):37-47. https://doi.org/10.4041/kjod24.134

Received: June 27, 2024; Revised: September 30, 2024; Accepted: October 18, 2024

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Objective: This study aimed to compare the accuracy of Qlone, Magiscan, and 3dMD with that of direct anthropometry (DA). Methods: The study involved 41 patients. Sixteen facial landmarks, including six individual and five paired points, were marked on each participant’s face. Subsequently, 18 linear measurements were assessed using a 3dMD device (multicamera photogrammetry), Qlone, Magiscan smartphone applications (single-camera photogrammetry), and DA. The Qlone and Magiscan images were calibrated using a reference point 10 mm from the nasion during DA to ensure a 1:1 correspondence. Results: Concerning the precision of the digital methods compared to DA, the mean intraclass correlation coefficient values of 3dMD, Qlone and Magiscan were 0.989, 0.980 and 0.982, respectively. Compared with DA, 3dMD achieved excellent trueness with the lowest average absolute differences in the measurements (highest value = 0.95 ± 0.62 mm). The highest values for Qlone and Magiscan were 1.51 ± 1.11 mm and 2.14 ± 1.69 mm, respectively. According to the number of parameters, the ranking of unreliable values (> 2 mm) was Magiscan (n = 46), Qlone (n = 35), and then, 3dMD (n = 4). Furthermore, reliability (less than 1 mm) was the highest for 3dMD (n = 517), followed by Magiscan (n = 457), and then, Qlone (n = 415). Conclusions: The 3dMD achieved excellent trueness with the lowest average absolute differences in the measurements. Based on statistical analysis, the trueness values of Magiscan and Qlone were close to that of 3dMD. To apply these smartphone applications clinically, more studies are necessary.

Keywords: Photography, 3-dimensional diagnosis, 3D scanner, Soft tissue

INTRODUCTION

Over the past few decades, facial esthetic analysis has gained significant importance in orthodontics. Orthodontists frequently rely on photographs to aid in esthetic diagnosis and treatment outcome evaluation.1,2 Traditional 2D photographs lack the accuracy needed to capture 3D facial details that are critical to comprehensive assessment. Hence, 3D photogrammetry is used.3,4 Historically, direct anthropometry (DA) was the gold standard for facial measurements owing to its affordability and reliability. However, DA is time consuming and operator dependent. Furthermore, it causes patient discomfort and data storage difficulties.5 To enhance measurement reliability and mitigate procedural errors, dentistry has shifted to 3D digital evaluation using facial scanning devices in recent years.6 3D imaging has many applications including pre-treatment evaluation of craniofacial deformities, cleft lip and palate, orthognathic surgery, syndromes, and facial asymmetries,7 as well as soft tissue and orthodontic treatment outcome evaluations.

3D digital imaging has been used in medicine since 1922.8 Various 3D facial scanning technologies, such as laser scanning, structured light scanning, stereophotogrammetry (multi-camera photogrammetry), photogrammetry (single-camera photogrammetry), and dual-structured light scanning, have been employed.9 The accuracy of laser scanning is hindered by lengthy scanning periods that may result in motion artifacts. All these systems are non-invasive, accurate and reproducible. Most systems require calibration (except for dual-structured light).10 The 3dMDface system (3dMD; 3dMD Inc., Atlanta, GA, USA) is widely used. It incorporates passive and active stereophotogrammetry in 3D surface imaging systems.11,12 Stereophotogrammetry offers several advantages. For example, the results are independent of light intensity and only one scan is required. 3dMD involves a multiple-camera system consisting of two modules, each with six cameras, and the shooting process does not require image stitching. The ability to capture an image in one scan prevents problems caused by patient movement.13 The disadvantages include high cost, large physical size, immobility, and its need for daily calibration.10,14

Most 3D photogrammetry systems, such as 3dMD and VECTRA M3 (3D Imaging System; Canfield Scientific, Parsippany, NJ, USA) are static devices. Currently, portable 3D imaging devices, such as VECTRA H1 and H2 (Canfield Scientific) comprise one digital single-lens reflex camera and a computer system. Typically, three consecutive photographs are taken from three angles, and then, merged to produce one 3D photo using computer software. Unlike static devices, portable instruments are relatively inexpensive and access is not location dependent. However, the reliability of portable devices is marginally lower than that of static devices.15,16

Photogrammetry is another widely used measurement technique. Although less expensive than stereophotogrammetry, photogrammetry requires several photographs and reverse engineering software for constructing a 3D image, while also being sensitive to light.10 More recently, cost-effective 3D scanning systems have emerged, including the dual-structured light system Bellus3D FaceMaker (Bellus 3D Inc., Campbell, CA, USA), which can be installed as an application on smartphones or tablets. This application facilitates 3D facial photography using an external camera integrated into a smartphone or tablet. Importantly, the accuracy of Bellus3D is similar to that of 3dMD, and the cost difference is almost negligible.14,17

Other smartphone 3D scanning applications include Qlone, Magiscan, EM3D, Heges, ScandyPro, Polycam, Widar, Kiri Engine, and Trnio.18 Of these, Qlone (version 5.6.0; EyeCue Vision Technologies, LTD-AR Technologies, Yokne’am, Israel) and Magiscan (version 1.5.12; AR Generation, AR and LiDAR Technologies, Warsaw, Poland) are the most user-friendly for clinical applications. Hence, we included these two applications into our study. To the best of our knowledge, the accuracy of these new applications has not been previously reported.

Magiscan and Qlone are 3D scanning applications that use single-camera photogrammetry and capture object distances using laser pulses, similar to a radar. An infrared transmitter on the mobile light detection and ranging (LiDAR) camera projects dots on the scanning surface,19 and thus, direct daylight impacts the scanning quality of LiDAR cameras.20 In contrast, Magiscan needs LiDAR sensors (compatible with iPhone 12 Pro and above), whereas Qlone employs photogrammetry without the need for LiDAR. Both applications use photogrammetry to convert a series of pictures taken on an iPhone or iPad into USDZ files that can be used to construct a 3D image or viewed in augmented reality. USDZ is Apple’s file format for augmented reality on the iOS platform. Meanwhile, the extended image acquisition time associated with these applications, compared with 3dMD, may lead to complications arising from patient movement. Assessment of currently available 3D photography applications is necessary to facilitate the integration of these novel methods into dentistry and orthodontics.

Deutsch et al.21 defined the anatomical landmarks for anthropometric measurements. In studies examining reliability, careful selection of landmarks is crucial, especially those affected by facial expressions (e.g., perioral points), with unclear visualization (e.g., tragus), or those that are difficult to detect anatomically (e.g., gonion). In the literature, landmarks located on the ears, such as tragions, are difficult to place precisely because of potential shadows and hair obstructions. The 3dMDface and Bellus 3D systems produce surface deviations around the ears; thus, the accuracy of ear landmarks is insufficient.16,22 If those areas must be measured, the 3dMD system may increase the precision of the landmarks on the ears.

The present study aimed to compare the accuracy of Qlone, Magiscan, and 3dMD, and DA. We hypothesized that photogrammetry systems on smartphones are reliable and cost-effective alternatives to stereophotogrammetry (3dMD).

MATERIALS AND METHODS

This single-center, single-blinded study was approved by the Ethical Committee of Marmara University Dentistry Faculty Clinical Researches Ethical Committee (approval date and protocol number: 26.01.2023 and 2023/121) and was conducted according to the Declaration of Helsinki (World Medical Association, 2013). Sample size calculations were performed using the G*Power program (version 3.1.9.2; Heinrich-Heine-University, Düsseldorf, Germany) with a statistical power of 90%, an alpha error of 0.05, and an effect size of 0.3. The sample population consisted of 41 patients, with a mean age of 24.74 ± 3.16 years. Participants were informed verbally and in writing before the study. Informed consent was obtained from all participants. The inclusion criteria were adults, students or staff of the School of Dentistry, aged 18–35 years, stable occlusion, willingness to participate, and signed informed consent. Individuals with dentofacial deformities, facial trauma, surgical scars, skin diseases, mustaches, beards, or acne covering facial landmarks were excluded.

Data collection

Initially, DA measurements were performed using the standard anthropometric landmarks defined by Deutsch et al.21 Of the 32 landmarks, 16 that were the least affected by gestures and movements, were selected. The selected landmarks (6 single and 5 paired) were as follows: reference point (ref), zygonion (zygo right [R]-left [L]), trichion (tri), nasion (n), subnasale (sn), Pronasale (prn), endocanthion (endo R-L), eyebrow (eb R-L), alare (ala R-L), pogonion (pog). These were initially marked on each individual’s face using an acetate pen with a 0.5 mm tip (Figure 1). Eighteen linear measurements based on the landmarks, including transverse, vertical, sagittal oblique, and vertical oblique measurements were evaluated (Table 1).

Figure 1. The landmarks used in the present study.
See Table 1 for definitions of each landmark or measurement.

Table 1 . Description of anthropometric landmarks and definition of linear distances.

DirectionAbbreviationDescription
TransversalendoR-endoLDistance between right and left endocanthions (inner corner of eye)
alaR-alaLDistance between right and left alares (most prominent point of alare)
ebR-ebLDistance between right and left initial point of eyebrows
zygoR-zygoLDistance between right and left zygonions (most prominent point in zygomatic region)
Sagittal-ObliquezygoL-alaLDistance between left zygonion and alare
zygoR-alaRDistance between right zygonion and alare
alaL-prnDistance between left alare and pronasale
alaR-prnDistance between right alare and pronasale
zygoL-prnDistance between left zygonion and pronasale
zygoR-prnDistance between right zygonion and pronasale
Verticaltri-nDistance between trichion and nasion
n-snDistance between nasion and subnasale
sn-pogDistance between subnasale and pogonion
n-prnDistance between nasion and pronasale
Vertical-ObliquezygoL-nDistance between left zygonion and nasion
zygoR-nDistance between right zygonion and nasion
zygoL-pogDistance between left zygonion and pogonion
zygoR-pogDistance between right zygonion and pogonion

endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..



The patients were seated on a height-adjustable chair and instructed to look at a mirror fixed in front of them to establish a natural head position. Images acquired from 3dMD were automatically saved in a tricorder surface binary (.tsb) format. Additionally, single-camera photogrammetry was performed using the smartphone applications, namely, Qlone and Magiscan. To construct the 3D images, Qlone and Magiscan captured 40 and 45 serial photographs per scan, respectively. 3D scans were obtained with the patient in a stable head position, natural expression, and maximum intercuspation while the camera moved around the head.17 All the scans were performed using a 48-megapixel True depth camera, which uses LiDAR, with an A16 Bionic chip (iPhone 14 Pro, iOS 16.1.1, Apple Store, Cupertino, CA, USA). After the scanning procedure, the 3D files were saved in the .obj format (Figure 2).

Figure 2. A, 3dMD image. B, Magiscan image. C, Qlone image.

Due to surface deviations around the ears, precisely placing landmarks in that region was difficult, resulting in insufficient accuracy. Consequently, the ear regions were removed from all images. To provide a visual representation, a sample color map was generated by superimposing three digital samples (Figure 3).

Figure 3. A 3D color map of the superimposed samples for a patient.

Data measurement

DA measurements were performed using a digital caliper (0–150 mm, Digital Caliper 1112-150, INSIZE Co., Ltd., Suzhou New District, China) with an accuracy of 0.03 mm. For the digital measurements, .obj files were imported into the free-source MeshLab software (ISTI [Italian National Research Council], Rome, Italy). A reference point was marked 10 mm above the nasion to calibrate the scanned photographs (Figure 1). The mean differences between the Qlone and Magiscan images and the DA were scaled using coefficients calculated based on a 10 mm scale, resulting in a 1:1 correspondence. Of note, the magnification factor for Magiscan was approximately 200, and approximately 1.1 to 1.3 times for Qlone. Therefore, a minor error may be amplified during the initial measurement owing to the multiplication of the coefficient.

All measurements were repeated on the same day by the same operator to calculate intra-examiner repeatability. Accuracy is the agreement between a measurement and the “true” value of a parameter, as provided by the DA and measured using a caliper. The average absolute distance between the DA and digital inter-landmark distances was assessed to evaluate accuracy. Additionally, the mean values of the parameters for each measurement method were compared to assess their reliability.

Statistical analysis

IBM SPSS Statistics (version 22.0; IBM Corp., Armonk, NY, USA) was used for all the statistical analyses. The conformity of the parameters to the normal distribution was assessed using the Shapiro–Wilk test. To evaluate the study data, the intraclass correlation coefficient (ICC) was calculated to determine the reliability of the digital methods compared to DA. The Friedman test was used to evaluate the amount of deviation of the other scanners compared to the direct anthropometric measurements. Bonferroni correction was performed for pairwise comparison. Intra-examiner reliability was assessed using the ICC. Statistical significance was set at P < 0.05.

RESULTS

The intra-examiner reliability of DA, 3dMD, Qlone, and Magiscan ranged from 0.990 to 1.000, 0.997 to 1.000, 0.991 to 1.000, and 0.981 to 1.000, respectively, indicating a high level of agreement. The mean values of each parameter for each measurement method are presented in Table 2. Concerning the precision of the digital methods compared with DA, the mean ICC values of the 3dMD, Qlone, and Magiscan were 0.989, 0.980, and 0.982, respectively. The lowest ICC was 0.972 for 3dMD, 0.950, and 0.949 for Qlone (Table 2).

Table 2 . Descriptive statistics of the variables for each method.

DA3dMDQloneMagiscanDA-3dMDDA-QloneDA-Magiscan
Mean ± SDMean ± SDMean ± SDMean ± SDICC (95% CI)ICC (95% CI)ICC (95% CI)
Transversal
endoR-endoL (mm)28.83 ± 2.4828.97 ± 2.5628.92 ± 2.5728.93 ± 2.380.976 (0.956–0.987)0.960 (0.925–0.979)0.984 (0.970–0.991)
alaR-alaL (mm)32.70 ± 2.3932.91 ± 2.5133.33 ± 2.6432.89 ± 2.450.972 (0.948–0.985)0.949 (0.904–0.973)0.978 (0.959–0.988)
ebR-ebL (mm)20.47 ± 4.3520.23 ± 4.2420.27 ± 4.4020.28 ± 4.240.995 (0.991–0.977)0.995 (0.991–0.998)0.994 (0.989–0.997)
zygoR-zygoL (mm)121.70 ± 7.80121.42 ± 7.79121.3 ± 7.97120.01 ± 7.780.996 (0.992–0.998)0.986 (0.974–0.993)0.980 (0.963–0.990)
Sagittal-Oblique
zygoR-alaR (mm)55.16 ± 6.1954.60 ± 6.1154.68 ± 6.1554.50 ± 6.130.997 (0.993–0.998)0.992 (0.985–0.996)0.992 (0.985–0.996)
zygoL-alaL (mm)59.35 ± 4.8058.92 ± 4.9859.08 ± 4.9958.74 ± 5.050.993 (0.987–0.996)0.998 (0.997–0.994)0.989 (0.979–0.994)
alaR-prn (mm)27.26 ± 1.9226.40 ± 1.7826.95 ± 1.9526.80 ± 1.890.974 (0.951–0.986)0.958 (0.921–0.977)0.959 (0.922–0.978)
alaL-prn (mm)27.03 ± 1.9026.39 ± 1.8826.74 ± 2.0626.69 ± 1.960.972 (0.948–0.985)0.962 (0.929–0.980)0.950 (0.906–0.973)
zygoR-prn (mm)79.53 ± 7.0978.61 ± 7.0379.25 ± 7.1978.47 ± 6.870.997 (0.995–0.998)0.991 (0.983–0.995)0.993 (0.986–0.996)
zygoL-prn (mm)83.75 ± 5.7082.91 ± 5.7383.27 ± 5.9382.66 ± 5.820.996 (0.992–0.998)0.982 (0.966–0.990)0.983 (0.968–0.991)
Vertical
tri-n (mm)63.15 ± 7.5063.40 ± 7.4663.28 ± 7.4062.86 ± 7.550.996 (0.993–0.998)0.993 (0.987–0.996)0.992 (0.986–0.996)
n-sn (mm)51.37 ± 3.6351.63 ± 3.5351.63 ± 3.6051.27 ± 3.460.987 (0.975–0.993)0.977 (0.957–0.998)0.981 (0.965–0.990)
sn-pog (mm)50.68 ± 4.2551.17 ± 4.3851.23 ± 4.3250.92 ± 4.260.989 (0.979–0.994)0.980 (0.963–0.989)0.984 (0.969–0.991)
n-prn (mm)43.10 ± 4.3943.19 ± 4.2743.37 ± 4.5042.90 ± 4.300.996 (0.992–0.998)0.989 (0.980–0.994)0.992 (0.985–0.996)
Vertical-Oblique
zygoR-n (mm)73.76 ± 5.4673.26 ± 5.4773.58 ± 5.4573.07 ± 5.450.993 (0.986–0.996)0.987 (0.976–0.993)0.986 (0.973–0.992)
zygoL-n (mm)75.05 ± 4.5774.57 ± 4.5675.00 ± 4.6274.34 ± 4.630.994 (0.998–0.997)0.981 (0.965–0.990)0.978 (0.959–0.988)
zygoR-pog (mm)96.98 ± 6.4496.86 ± 6.4097.09 ± 6.8396.19 ± 6.600.996 (0.992–0.998)0.989 (0.980–0.994)0.989 (0.980–0.994)
zygoL-pog (mm)100.56 ± 6.13100.95 ± 6.34100.82 ± 6.34100.02 ± 6.100.995 (0.991–0.997)0.981 (0.964–0.990)0.985 (0.971–0.992)

ICC, intraclass correlation coefficient; DA, direct anthropometry; SD, standard deviation; CI, confidence interval; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..



Table 3 presents the average absolute differences in the linear distances between the DA and digital methods. Compared with DA (the reference standard for the present study), 3dMD achieved high accuracy and the lowest average absolute difference, with the highest value being 0.95 ± 0.62 mm. Conversely, Qlone and Magiscan showed greater differences relative to DA. Moreover, Qlone showed significantly higher average absolute differences in the parameters of endoR-endoL (0.80 ± 0.59 mm), alaR-alaL (1.05 ± 0.72 mm), n-prn (0.77 ± 0.57 mm), zygoL-n (1.08 ± 0.60 mm), and zygoL-pog (1.42 ± 0.96 mm), whereas the highest significant value in Magiscan was observed for zygoR-zygoL (2.14 ± 1.69 mm) (P < 0.05).

Table 3 . Average absolute differences between DA and other scanning methods and pairwise comparisons of the average absolute differences between the methods.

ParameterDA-3dMDDA-QloneDA-MagiscanPa3dMD-Qlone
(Pb)
3dMD-Magiscan
(Pb)
Qlone-Magiscan
(Pb)
Mean ± SDMean ± SDMean ± SD
Transversal
endoR-endoL (mm)0.61 ± 0.480.80 ± 0.590.47 ± 0.400.049*1.0000.3660.045*
alaR-alaL (mm)0.67 ± 0.481.05 ± 0.720.44 ± 0.570.000***0.0610.0610.000***
ebR-ebL (mm)0.49 ± 0.410.43 ± 0.450.51 ± 0.430.438---
zygoR-zygoL (mm)0.89 ± 0.541.51 ± 1.112.14 ± 1.690.001***0.0820.001***0.502
Sagittal-Oblique
zygoR-alaR (mm)0.79 ± 0.481.00 ± 0.671.05 ± 0.700.109---
zygoL-alaL (mm)0.75 ± 0.510.89 ± 0.650.98 ± 0.700.181---
alaR-prn (mm)0.89 ± 0.540.63 ± 0.540.74 ± 0.490.071---
alaL-prn (mm)0.67 ± 0.580.65 ± 0.470.76 ± 0.490.426---
zygoR-prn (mm)0.95 ± 0.621.15 ± 0.771.32 ± 0.880.787---
zygoL-prn (mm)0.93 ± 0.611.20 ± 1.081.45 ± 1.130.067---
Vertical
tri-n (mm)0.77 ± 0.520.94 ± 0.771.06 ± 0.810.249---
n-sn (mm)0.70 ± 0.500.93 ± 0.610.81 ± 0.510.081---
sn-pog (mm)0.92 ± 0.441.05 ± 0.780.93 ± 0.590.843---
n-prn (mm)0.44 ± 0.370.77 ± 0.570.65 ± 0.470.028*0.024*0.9610.293
Vertical-Oblique
zygoR-n (mm)0.88 ± 0.591.08 ± 0.591.06 ± 0.990.232---
zygoL-n (mm)0.70 ± 0.521.08 ± 0.601.16 ± 0.980.021*0.017*0.3660.673
zygoR-pog (mm)0.67 ± 0.491.08 ± 0.801.17 ± 1.030.157---
zygoL-pog (mm)0.78 ± 0.541.42 ± 0.961.13 ± 1.120.002**0.002**0.9610.039*

DA, direct anthropometry; SD, standard deviation; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..

aFriedman test, bBonferroni correction for multiple tests..

*P < 0.05, **P < 0.01, ***P < 0.001..



The mean absolute differences of the linear measurements for 3dMD, Qlone, and Magiscan were 0.75 ± 0.53 mm, 0.98 ± 0.77 mm, and 0.99 ± 0.92 mm, respectively. Figure 4 shows the reliability of the methods according to the average absolute differences for all the parameters (738 in total). Notably, 3dMD demonstrated four unreliable measurements, while Qlone and Mgiscan showed 35 and 46, respectively.

Figure 4. Reliability differences among 3dMD, Magiscan, and Qlone.

DISCUSSION

This study compared the accuracy of the Qlone, Magiscan, and 3dMD systems to each other and to the DA method. During the preliminary stages of the research design, multiple smartphone applications were evaluated, among which EM3D, Heges, and ScandyPro generated 3D images by employing a video-recording approach analogous to that of Bellus 3D. EM3D and Heges produced images using a lower-megapixel front camera, while ScandyPro used the rear camera.18 All of these applications demonstrated challenges regarding accurate image capture and tracking loss. Considering that 3D scanning technologies depend on photogrammetry, the 3D images created in Magiscan and Qlone by combining serial photos taken using a smartphone’s high-megapixel rear camera produced better image quality for the construction of 3D photographs without tracking loss. Moreover, the selected applications were the most cost efficient.

Based on empirical observations, optimal 3D image acquisition with those applications required a well-illuminated environment with direct daylight exposure on the participant’s face.18 This divergence from 3dMD’s performance characteristics may be attributed to the light sensitivity of the applications.10

In the present study, angular measurements were excluded because of the inherent challenges associated with DA stemming from operator experience and soft tissue resilience. Additionally, changes in facial expressions may affect the positions of landmarks in the perioral region, and it is difficult to precisely locate the gonion without direct palpation. Therefore, these landmarks were excluded from the study.

For the ICC, a value greater than 0.900 indicated that the measurements were consistent and reproducible. Most previous studies reported ICC values higher than 0.900 for 3dMD,5,17,23,24 Dental Pro,25 Face Hunter,25 and Bellus 3D.17,23,24 Comparing the DA and digital methods, all measurements showed high agreement, with ICC values greater than 0.900. Among the evaluated methods, 3dMD presented the highest ICC values for most parameters, followed by Magiscan, and then Qlone. The lowest ICC value was 0.972 for 3dMD, 0.950 for Magiscan, and 0.949 for Qlone. These results indicate that each method was highly reliable and precise compared with DA.

Liu et al.17 reported a mean absolute difference of 0.36 ± 0.20 mm for 3dMD. In the present study, the clinical acceptance has been set at 1 mm. The accuracy of 3dMD was found to be clinically acceptable5 and showed a mean absolute difference of 0.75 ± 0.53 mm compared with DA. Additionally, Qlone and Magiscan achieved mean absolute differences of 0.98 ± 0.77 mm and 0.99 ± 0.92 mm, respectively, which are considered generally reliable and within acceptable limits (except for the zygonion measurements). This may be explained by the lateral positioning of the zygo to the midline, which is located on the curvature of the face, and the enlargement of that area with these techniques. Considering the relatively low accuracy associated with the zygonion landmark for these two methods, the exocanthion may be appropriate in future oblique measurement studies. However, the risk of squinting and displacing these points should not be ignored, especially in methods in which image acquisition takes a long time. In a previous study of the Dental Pro and Face Hunter applications, significant changes of the zygo R-L measurement was reported.25 Although similar differences were not found to be significant in the present study, differences over 1 mm were observed in the measurements including zygo landmarks, in the Magiscan and Qlone 3D images.

Additionally, the measurement that included the sn point showed a 0.70–1.05 mm difference using the three digital imaging techniques compared with DA. This is because the sn is difficult to locate due to the curvature and positional variations of the lip (including lip posture and morphology) and the nose.26 In 3D imaging, attention should be paid to parameters involving the sn. To summarize, larger differences were observed when acquiring areas with more curvature, especially when using Qlone and Magiscan. These results are supported by Hong et al.,27 who reported the greatest errors in areas with curvatures and larger variations, especially in longer measurements. Additional improvements, especially in smartphone applications, may be necessary to accurately capture the areas of curvature.

Dindaroğlu et al.5 reported a deviation ranging from −0.21 to 0.15 mm between DA and 3dMD. However, the deviation values were not average absolute differences, such as those calculated in the present study.5 This may be the reason why the results reported by Dindaroğlu et al.5 were lower than the results presented herein. Dindaroğlu et al.5 found a significant change in the endo R-L distance (ranging between −1.16 and 0.74 mm) when DA and 3dMD were compared. In the present study, this parameter was significant when the average absolute differences of Qlone (0.80 mm) and Magiscan (0.47 mm) compared to DA were considered, showing higher differences in Qlone.

Raffone et al.28 found the difference between the ala to be 0.788 mm for Bellus 3D. In this study, the ala R-L measured 0.67 mm for 3dMD, 1.05 mm for Qlone, and 0.44 mm for Magiscan. Additionally, the differences in the ala-prn distance between Bellus 3D and DA was 0.524 mm on the right side and 1.177 mm on the left side.28 The differences for those parameters in the present study were 0.89 mm for 3dMD, 0.63 mm for Qlone, and 0.74 mm for Magiscan on the right side; 0.67 mm for 3dMD, 0.65 mm for Qlone, and 0.76 mm for Magiscan on the left side. Moreover, the alar wings moved differently, depending on the breathing pattern during image acquisition. For instance, the alar wings of a nasal breather tended to expand during breathing, whereas an oral breather presents with narrowing or inactivity of the alar wings. This may help explain the large ala measurement differences in studies.

The average absolute difference in the n-prn between DA and 3dMD was found to be 0.44 mm, which was consistent with the results of previous reports (0.3–0.4 mm).17 Furthermore, Qlone and Magiscan exhibited a deviation of 0.77 mm and 0.65 mm, respectively. Raffone et al.28 compared the slider and free techniques. With the slider technique, the patient’s head is motionless and the scanner moves around the head. With the free technique, head is rotated, while the scanner remains stable. They reported that the slider technique was more suitable for clinical use due to fewer motion artifacts, thus, resulting in a scan free from distortions.28 In the present study, the images were obtained by employing the sliding technique.

The ICCs for the agreement between Bellus 3D and DA mostly fell below the acceptable limit of 0.80.24 In contrast, all the methods were found to be highly precise in the present study, presenting high ICC values compared with DA. According to the results, Magiscan and Qlone exhibited fewer differences than Bellus 3D (dual-structure light). However, studies that include the Bellus 3D and smartphone applications may provide more definitive and reliable results.

To clarify the results, a reliability table was constructed by employing a 2 mm reliability threshold. This categorization, referring to established studies, classified the measurements into four intervals.29 Although some studies considered a 2 mm discrepancy as clinically acceptable,28 others proposed a threshold of 1 mm.5,17 However, developing technology should promise to reduce this limit. Thus, we adopted a 1 mm deviation as the criterion for clinical acceptability.

This present study included 738 measurements. Of these, 76.28% and 81.43% of the Magiscan and Qlone measurements provided highly reliable outcomes, with 6.23% and 4.74% for Magiscan and Qlone, respectively. Comparable analyses in the literature found 84% of the measurements to be highly reliable and reliable, 5% to be unreliable using Dental Pro, 87% to be highly reliable and reliable, and 1% to be unreliable using Face Hunter.25

The post-capture image-processing time with photogrammetry in Magiscan (5–10 minutes) was notably longer than that in Qlone (1–2 minutes), which is a disadvantage for clinical application. Both the Magiscan and Qlone require environments with sufficient daylight. Finally, the possibility of deformation in pointed and curved areas of the face, such as the zygo and alar wings, should be considered. To date, available studies are limited and some have performed their analyses using a mannequin head.10,17,25,27,28 Since the mannequin head is not affected by movement, methods must be tested in humans to accurately determine its clinical applicability. After all, measurements must ultimately be performed on individuals in the clinical setting.

One limitation of the present study is the absence of inter-examiner reliability. The operator’s smartphone handling skills may significantly impact the results. In the future, more studies on smartphone application updates and the different scanning techniques are needed, so that more people can create consistent 3D images using the same program.

CONCLUSIONS

Qlone presented significantly higher values for the parameters n-prn, zygoL-n, and zygoL-pog, while the highest significant value for Magiscan was observed for ZygoR-zygoL compared with 3dMD.

The 3dMD achieved high accuracy and exhibited the lowest average absolute differences. Magiscan and QLONE are cost-effective alternatives to 3dMD. Hence, their integration into clinical routines may be recommended without ignoring the fact that Magiscan and Qlone have 10 times the number of unreliable measurements compared with 3dMD.

The absence of supplementary cameras and their economical pricing render Qlone and Magiscan advantageous compared to alternative systems. Further studies are required to validate clinical reliability.

The post-capture image processing time associated with photogrammetry using Magiscan was notably longer than that using Qlone, which is a disadvantage of 3dMD.

AUTHOR CONTRIBUTIONS

Conceptualization: GK, HNY. Data curation: GY. Formal analysis: GY. Investigation: GK. Methodology: GK, HNY. Project administration: GK, HNY. Resources: HNY. Software: GY. Supervision: GK, HNY. Validation: HNY, GY. Visualization: GK, GY. Writing–original draft: GK. Writing–review & editing: HNY.

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

FUNDING

None to declare.

Fig 1.

Figure 1.The landmarks used in the present study.
See Table 1 for definitions of each landmark or measurement.
Korean Journal of Orthodontics 2025; 55: 37-47https://doi.org/10.4041/kjod24.134

Fig 2.

Figure 2.A, 3dMD image. B, Magiscan image. C, Qlone image.
Korean Journal of Orthodontics 2025; 55: 37-47https://doi.org/10.4041/kjod24.134

Fig 3.

Figure 3.A 3D color map of the superimposed samples for a patient.
Korean Journal of Orthodontics 2025; 55: 37-47https://doi.org/10.4041/kjod24.134

Fig 4.

Figure 4.Reliability differences among 3dMD, Magiscan, and Qlone.
Korean Journal of Orthodontics 2025; 55: 37-47https://doi.org/10.4041/kjod24.134

Table 1 . Description of anthropometric landmarks and definition of linear distances.

DirectionAbbreviationDescription
TransversalendoR-endoLDistance between right and left endocanthions (inner corner of eye)
alaR-alaLDistance between right and left alares (most prominent point of alare)
ebR-ebLDistance between right and left initial point of eyebrows
zygoR-zygoLDistance between right and left zygonions (most prominent point in zygomatic region)
Sagittal-ObliquezygoL-alaLDistance between left zygonion and alare
zygoR-alaRDistance between right zygonion and alare
alaL-prnDistance between left alare and pronasale
alaR-prnDistance between right alare and pronasale
zygoL-prnDistance between left zygonion and pronasale
zygoR-prnDistance between right zygonion and pronasale
Verticaltri-nDistance between trichion and nasion
n-snDistance between nasion and subnasale
sn-pogDistance between subnasale and pogonion
n-prnDistance between nasion and pronasale
Vertical-ObliquezygoL-nDistance between left zygonion and nasion
zygoR-nDistance between right zygonion and nasion
zygoL-pogDistance between left zygonion and pogonion
zygoR-pogDistance between right zygonion and pogonion

endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..


Table 2 . Descriptive statistics of the variables for each method.

DA3dMDQloneMagiscanDA-3dMDDA-QloneDA-Magiscan
Mean ± SDMean ± SDMean ± SDMean ± SDICC (95% CI)ICC (95% CI)ICC (95% CI)
Transversal
endoR-endoL (mm)28.83 ± 2.4828.97 ± 2.5628.92 ± 2.5728.93 ± 2.380.976 (0.956–0.987)0.960 (0.925–0.979)0.984 (0.970–0.991)
alaR-alaL (mm)32.70 ± 2.3932.91 ± 2.5133.33 ± 2.6432.89 ± 2.450.972 (0.948–0.985)0.949 (0.904–0.973)0.978 (0.959–0.988)
ebR-ebL (mm)20.47 ± 4.3520.23 ± 4.2420.27 ± 4.4020.28 ± 4.240.995 (0.991–0.977)0.995 (0.991–0.998)0.994 (0.989–0.997)
zygoR-zygoL (mm)121.70 ± 7.80121.42 ± 7.79121.3 ± 7.97120.01 ± 7.780.996 (0.992–0.998)0.986 (0.974–0.993)0.980 (0.963–0.990)
Sagittal-Oblique
zygoR-alaR (mm)55.16 ± 6.1954.60 ± 6.1154.68 ± 6.1554.50 ± 6.130.997 (0.993–0.998)0.992 (0.985–0.996)0.992 (0.985–0.996)
zygoL-alaL (mm)59.35 ± 4.8058.92 ± 4.9859.08 ± 4.9958.74 ± 5.050.993 (0.987–0.996)0.998 (0.997–0.994)0.989 (0.979–0.994)
alaR-prn (mm)27.26 ± 1.9226.40 ± 1.7826.95 ± 1.9526.80 ± 1.890.974 (0.951–0.986)0.958 (0.921–0.977)0.959 (0.922–0.978)
alaL-prn (mm)27.03 ± 1.9026.39 ± 1.8826.74 ± 2.0626.69 ± 1.960.972 (0.948–0.985)0.962 (0.929–0.980)0.950 (0.906–0.973)
zygoR-prn (mm)79.53 ± 7.0978.61 ± 7.0379.25 ± 7.1978.47 ± 6.870.997 (0.995–0.998)0.991 (0.983–0.995)0.993 (0.986–0.996)
zygoL-prn (mm)83.75 ± 5.7082.91 ± 5.7383.27 ± 5.9382.66 ± 5.820.996 (0.992–0.998)0.982 (0.966–0.990)0.983 (0.968–0.991)
Vertical
tri-n (mm)63.15 ± 7.5063.40 ± 7.4663.28 ± 7.4062.86 ± 7.550.996 (0.993–0.998)0.993 (0.987–0.996)0.992 (0.986–0.996)
n-sn (mm)51.37 ± 3.6351.63 ± 3.5351.63 ± 3.6051.27 ± 3.460.987 (0.975–0.993)0.977 (0.957–0.998)0.981 (0.965–0.990)
sn-pog (mm)50.68 ± 4.2551.17 ± 4.3851.23 ± 4.3250.92 ± 4.260.989 (0.979–0.994)0.980 (0.963–0.989)0.984 (0.969–0.991)
n-prn (mm)43.10 ± 4.3943.19 ± 4.2743.37 ± 4.5042.90 ± 4.300.996 (0.992–0.998)0.989 (0.980–0.994)0.992 (0.985–0.996)
Vertical-Oblique
zygoR-n (mm)73.76 ± 5.4673.26 ± 5.4773.58 ± 5.4573.07 ± 5.450.993 (0.986–0.996)0.987 (0.976–0.993)0.986 (0.973–0.992)
zygoL-n (mm)75.05 ± 4.5774.57 ± 4.5675.00 ± 4.6274.34 ± 4.630.994 (0.998–0.997)0.981 (0.965–0.990)0.978 (0.959–0.988)
zygoR-pog (mm)96.98 ± 6.4496.86 ± 6.4097.09 ± 6.8396.19 ± 6.600.996 (0.992–0.998)0.989 (0.980–0.994)0.989 (0.980–0.994)
zygoL-pog (mm)100.56 ± 6.13100.95 ± 6.34100.82 ± 6.34100.02 ± 6.100.995 (0.991–0.997)0.981 (0.964–0.990)0.985 (0.971–0.992)

ICC, intraclass correlation coefficient; DA, direct anthropometry; SD, standard deviation; CI, confidence interval; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..


Table 3 . Average absolute differences between DA and other scanning methods and pairwise comparisons of the average absolute differences between the methods.

ParameterDA-3dMDDA-QloneDA-MagiscanPa3dMD-Qlone
(Pb)
3dMD-Magiscan
(Pb)
Qlone-Magiscan
(Pb)
Mean ± SDMean ± SDMean ± SD
Transversal
endoR-endoL (mm)0.61 ± 0.480.80 ± 0.590.47 ± 0.400.049*1.0000.3660.045*
alaR-alaL (mm)0.67 ± 0.481.05 ± 0.720.44 ± 0.570.000***0.0610.0610.000***
ebR-ebL (mm)0.49 ± 0.410.43 ± 0.450.51 ± 0.430.438---
zygoR-zygoL (mm)0.89 ± 0.541.51 ± 1.112.14 ± 1.690.001***0.0820.001***0.502
Sagittal-Oblique
zygoR-alaR (mm)0.79 ± 0.481.00 ± 0.671.05 ± 0.700.109---
zygoL-alaL (mm)0.75 ± 0.510.89 ± 0.650.98 ± 0.700.181---
alaR-prn (mm)0.89 ± 0.540.63 ± 0.540.74 ± 0.490.071---
alaL-prn (mm)0.67 ± 0.580.65 ± 0.470.76 ± 0.490.426---
zygoR-prn (mm)0.95 ± 0.621.15 ± 0.771.32 ± 0.880.787---
zygoL-prn (mm)0.93 ± 0.611.20 ± 1.081.45 ± 1.130.067---
Vertical
tri-n (mm)0.77 ± 0.520.94 ± 0.771.06 ± 0.810.249---
n-sn (mm)0.70 ± 0.500.93 ± 0.610.81 ± 0.510.081---
sn-pog (mm)0.92 ± 0.441.05 ± 0.780.93 ± 0.590.843---
n-prn (mm)0.44 ± 0.370.77 ± 0.570.65 ± 0.470.028*0.024*0.9610.293
Vertical-Oblique
zygoR-n (mm)0.88 ± 0.591.08 ± 0.591.06 ± 0.990.232---
zygoL-n (mm)0.70 ± 0.521.08 ± 0.601.16 ± 0.980.021*0.017*0.3660.673
zygoR-pog (mm)0.67 ± 0.491.08 ± 0.801.17 ± 1.030.157---
zygoL-pog (mm)0.78 ± 0.541.42 ± 0.961.13 ± 1.120.002**0.002**0.9610.039*

DA, direct anthropometry; SD, standard deviation; endo, endochantion; ala, alare; eb, eyebrow; zygo, zygonion; prn, pronasale; tri, trichion; n, nasion; sn, subnasale; pog, pogonion; R, right; L, left..

aFriedman test, bBonferroni correction for multiple tests..

*P < 0.05, **P < 0.01, ***P < 0.001..


References

  1. Nair AG, Santhanam A. Clinical photography for periorbital and facial aesthetic practice. J Cutan Aesthet Surg 2016;9:115-21. https://doi.org/10.4103/0974-2077.184047
    Pubmed KoreaMed CrossRef
  2. Stebel A, Desmedt D, Bronkhorst E, Kuijpers MA, Fudalej PS. Rating nasolabial appearance on three-dimensional images in cleft lip and palate: a comparison with standard photographs. Eur J Orthod 2016;38:197-201. https://doi.org/10.1093/ejo/cjv024
    Pubmed KoreaMed CrossRef
  3. Jyothikiran H, Shanthara JR, Subbiah P, Thomas M. Craniofacial imaging in orthodontics--past present and future. Int J Orthod Milwaukee 2014;25:21-6. https://pubmed.ncbi.nlm.nih.gov/24812737/
  4. Da Silveira AC, Daw JL Jr, Kusnoto B, Evans C, Cohen M. Craniofacial applications of three-dimensional laser surface scanning. J Craniofac Surg 2003;14:449-56. https://doi.org/10.1097/00001665-200307000-00009
    Pubmed CrossRef
  5. Dindaroğlu F, Kutlu P, Duran GS, Görgülü S, Aslan E. Accuracy and reliability of 3D stereophotogrammetry: a comparison to direct anthropometry and 2D photogrammetry. Angle Orthod 2016;86:487-94. https://doi.org/10.2319/041415-244.1
    Pubmed KoreaMed CrossRef
  6. Li Q, Bi M, Yang K, Liu W. The creation of a virtual dental patient with dynamic occlusion and its application in esthetic dentistry. J Prosthet Dent 2021;126:14-8. https://doi.org/10.1016/j.prosdent.2020.08.026
    Pubmed CrossRef
  7. Erten O, Yılmaz BN. Three-dimensional imaging in orthodontics. Turk J Orthod 2018;31:86-94. https://doi.org/10.5152/TurkJOrthod.2018.17041
    Pubmed KoreaMed CrossRef
  8. Mannsbach M. Die stereophotogrammetrie als hilfsmittel in der orthodontie. Dtsch Zahnärzt Orthop 1922;14:105. https://scholar.google.co.kr/scholar?hl=ko&as_sdt=0%2C5&q=%22Die+stereophotogrammetrie+als+hilfsmittel+in+der+orthodontie%22&btnG=
  9. Magnani M, Douglass M. Photogrammetry and stereophotogrammetry. In: López Varela SL, ed. The encyclopedia of archaeological sciences. Chichester: Wiley Blackwell; 2019. p. 1-4. https://doi.org/10.1002/9781119188230.saseas0451
    CrossRef
  10. Piedra-Cascón W, Meyer MJ, Methani MM, Revilla-León M. Accuracy (trueness and precision) of a dual-structured light facial scanner and interexaminer reliability. J Prosthet Dent 2020;124:567-74. https://doi.org/10.1016/j.prosdent.2019.10.010
    Pubmed CrossRef
  11. Yüksel Coşkun E, Esenlik E. A prospective study comparing adolescent and post-adolescent periods regarding effects of activator appliance in patients with class II mandibular retrognathia by using 3dMDface analysis and cephalometry. Med Sci Monit 2020;26:e921401. https://doi.org/10.12659/msm.921401
    Pubmed KoreaMed CrossRef
  12. White JD, Ortega-Castrillon A, Virgo C, Indencleef K, Hoskens H, Shriver MD, et al. Sources of variation in the 3dMDface and Vectra H1 3D facial imaging systems. Sci Rep 2020;10:4443. https://doi.org/10.1038/s41598-020-61333-3
    Pubmed KoreaMed CrossRef
  13. Zhao YJ, Xiong YX, Wang Y. Three-dimensional accuracy of facial scan for facial deformities in clinics: a new evaluation method for facial scanner accuracy. PLoS One 2017;12:e0169402. https://doi.org/10.1371/journal.pone.0169402
    Pubmed KoreaMed CrossRef
  14. Aynechi N, Larson BE, Leon-Salazar V, Beiraghi S. Accuracy and precision of a 3D anthropometric facial analysis with and without landmark labeling before image acquisition. Angle Orthod 2011;81:245-52. https://doi.org/10.2319/041810-210.1
    Pubmed KoreaMed CrossRef
  15. Fan W, Guo Y, Hou X, Liu J, Li S, Ju S, et al. Validation of the portable next-generation VECTRA H2 3D imaging system for periocular anthropometry. Front Med (Lausanne) 2022;9:833487. https://doi.org/10.3389/fmed.2022.833487
    Pubmed KoreaMed CrossRef
  16. Loy RCH, Liew MKM, Yong CW, Wong RCW. Validation of low-cost mobile phone applications and comparison with professional imaging systems for three-dimensional facial imaging: a pilot study. J Dent 2023;137:104676. https://doi.org/10.1016/j.jdent.2023.104676
    Pubmed CrossRef
  17. Liu J, Zhang C, Cai R, Yao Y, Zhao Z, Liao W. Accuracy of 3-dimensional stereophotogrammetry: comparison of the 3dMD and Bellus3D facial scanning systems with one another and with direct anthropometry. Am J Orthod Dentofacial Orthop 2021;160:862-71. https://doi.org/10.1016/j.ajodo.2021.04.020
    Pubmed CrossRef
  18. Van Lint L, Christiaens L, Stroo V, Bila M, Willaert R, Sun Y, et al. Accuracy comparison of 3D face scans obtained by portable stereophotogrammetry and smartphone applications. J Med Biol Eng 2023;43:550-60. https://doi.org/10.1007/s40846-023-00817-9
    CrossRef
  19. Harrap R, Daniel S. Mobile lidar mapping: building the next generation of outdoor environment models for AR [Internet]. IEEE International Symposium on Mixed and Augmented Reality ISMAR: Orlando, FL; 2009 [cited 2009 Oct 19]. Available from: https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=7406fce8f7fc17a083eb533aafae2b91055d3961
  20. Akan B, Akan E, Şahan AO, Kalak M. Evaluation of 3D face-scan images obtained by stereophotogrammetry and smartphone camera. Int Orthod 2021;19:669-78. https://doi.org/10.1016/j.ortho.2021.08.007
    Pubmed CrossRef
  21. Deutsch CK, Shell AR, Francis RW, Bird BD. The farkas system of craniofacial anthropometry: methodology and normative databases. In: Preedy VR, ed. Handbook of anthropometry: physical measures of human form in health and disease. New York: Springer New York; 2012. p. 561-73. https://doi.org/10.1007/978-1-4419-1788-1_29
    CrossRef
  22. Aldridge K, Boyadjiev SA, Capone GT, DeLeon VB, Richtsmeier JT. Precision and error of three-dimensional phenotypic measures acquired from 3dMD photogrammetric images. Am J Med Genet A 2005;138a:247-53. https://doi.org/10.1002/ajmg.a.30959
    Pubmed KoreaMed CrossRef
  23. Weinberg SM, Naidoo S, Govier DP, Martin RA, Kane AA, Marazita ML. Anthropometric precision and accuracy of digital three-dimensional photogrammetry: comparing the Genex and 3dMD imaging systems with one another and with direct anthropometry. J Craniofac Surg 2006;17:477-83. https://doi.org/10.1097/00001665-200605000-00015
    Pubmed CrossRef
  24. Staller S, Anigbo J, Stewart K, Dutra V, Turkkahraman H. Precision and accuracy assessment of single and multicamera three-dimensional photogrammetry compared with direct anthropometry. Angle Orthod 2022;92:635-41. https://doi.org/10.2319/101321-770.1
    Pubmed KoreaMed CrossRef
  25. Pellitteri F, Brucculeri L, Spedicato GA, Siciliani G, Lombardo L. Comparison of the accuracy of digital face scans obtained by two different scanners. Angle Orthod 2021;91:641-9. https://doi.org/10.2319/092720-823.1
    Pubmed KoreaMed CrossRef
  26. Masoud MI, Bansal N, Castillo JC, Manosudprasit A, Allareddy V, Haghi A, et al. 3D dentofacial photogrammetry reference values: a novel approach to orthodontic diagnosis. Eur J Orthod 2017;39:215-25. https://doi.org/10.1093/ejo/cjw055
    Pubmed CrossRef
  27. Hong C, Choi K, Kachroo Y, Kwon T, Nguyen A, McComb R, et al. Evaluation of the 3dMDface system as a tool for soft tissue analysis. Orthod Craniofac Res 2017;20 Suppl 1:119-24. https://doi.org/10.1111/ocr.12178
    Pubmed KoreaMed CrossRef
  28. Raffone C, Gianfreda F, Bollero P, Pompeo MG, Miele G, Canullo L. Chairside virtual patient protocol. Part 1: free vs guided face scan protocol. J Dent 2022;116:103881. https://doi.org/10.1016/j.jdent.2021.103881
    Pubmed CrossRef
  29. Aung SC, Ngim RC, Lee ST. Evaluation of the laser scanner as a surface measuring tool and its accuracy compared with direct facial anthropometric measurements. Br J Plast Surg 1995;48:551-8. https://doi.org/10.1016/0007-1226(95)90043-8
    Pubmed CrossRef