Abstract
Aim:
Facial imaging technology has become a pivotal tool in modern medical practice, particularly within fields such as maxillofacial prosthodontics, orthodontics, and smile design. The creation of digital twins, or virtual patients, enhances diagnostic accuracy, aids in treatment planning, and improves outcome prediction. The aim of the study was to assess the accuracy of various facial scanners, determine overall accuracy of each scanner, and identify which scanner demonstrates superior accuracy in specific facial regions.
Settings and Design:
An observational crossover study.
Materials and Methods:
Cone beam computed tomography volumetric scan was used as a control group, as it has been considered as a gold standard in terms of accuracy. For comparison, scan data were obtained from three different scanners, namely Carestream facial scanner, Medit intraoral scanner for facial scan, and MetiSmile face scanner. The standard tessellation language files thus obtained were compared for accuracy in Geomagic X software by superimposition technique and were evaluated for their accuracy using various reference points on the face.
Statistical Analysis Used:
Normality was confirmed using the Shapiro–Wilk test. One-way analysis of variance for comparison among groups and Tukey test for pairwise comparison was used using SPSS software (IBM SPSS version 29 USA).
Results:
The study concluded that MetiSmile was the best facial scanner among the three groups with a mean discrepancy of (0.35 ± 0.33) mm and P = 0.001, indicating significant difference between the scanners.
Conclusion:
Each scanner evaluated demonstrated acceptable performance, with notable variations attributable to their distinct scanning methodologies. Among these, the MetiSmile scanner emerged as the most accurate, delivering the most favorable results in terms of accuracy.
Keywords: Accuracy, cone beam computed tomography, digital dentistry, facial scanners, scanning, Standard Tessellation Language
INTRODUCTION
Digital dentistry is revolutionizing the field by offering advanced, efficient, and highly precise alternatives to conventional methods. Oral and facial rehabilitation require thorough diagnosis and meticulous treatment planning.[1] Assessing facial morphology is crucial for diagnosing maxillofacial irregularities, smile designing, performing surgeries, crafting prosthesis, orthodontics, and evaluating postsurgery results. The use of three-dimensional (3D) facial models for patient simulations can significantly enhance treatment esthetics and outcome forecasting.[2,3] Traditionally, creating patient facial models involves a physical facial impression process, where elastomeric substances are applied, and a gypsum cast is made as a replica.[4] However, this approach can be uncomfortable for patients. In addition, the dimensional fidelity of these physical models is susceptible to various influencing factors, such as the viscosity of the impression material, setting times, storage conditions, and the delay between mixing the material and casting it with stone.[5] Moreover, the intricate structure of the human face poses a challenge for achieving an accurate and lifelike replication.[6] Traditionally, approaches for analyzing facial appearance have solely relied on two-dimensional (2D) measurement techniques, such as taking sequential 2D photographs from various angles and using various instruments like Vernier calipers and bevel protractors to assess the distances and angles.[7,8] Recent advancements in digital dentistry and optical scanning have shifted facial morphology research from 2D to more versatile 3D methods.[9] One of the gold standard 3D Scanning methods used is volumetric cone beam computed tomography (CBCT) scan, it is generally considered as the most accurate scan. However, due to certain disadvantages including cost, technique sensitivity, ionizing radiation, and nonportability, they are difficult to employ everywhere; this is where surface facial scanners come into play. Facial scanners employ four primary techniques: Photogrammetry, stereophotogrammetry, structured light scanning, and laser scanning. These techniques can be further categorized into two types of methodologies: passive and active. The passive methods involve capturing a person’s face using two or more photographs (as seen in photogrammetry and stereophotogrammetry). In contrast, active methods rely on 3D sensors to detect light patterns through processes like active triangulation, utilizing tools such as laser beams and structured light.[10] A 3D facial scanner is therefore an advanced noncontact optical measurement device capable of capturing 3D facial models in an open data format while preserving the real skin texture and color.[11] The scanning process is typically very rapid, often taking <1 s. Increasingly, studies have noted that 3D facial scanners can be integrated into dental clinics, where the 3D models produced can be applied for quantitative 3D diagnostics and treatment assessments.[12] A facial digitizer creates digital files in specific formats, such as standard tessellation language (STL), Object File Format (OBJ), Polygon File Format (PLY) files. Per ISO 5725-1, the concept of accuracy merges both trueness and precision. Trueness indicates how closely the scanner can replicate a 3D reconstruction that mirrors the original form. Precision, on the other hand, refers to the consistency and agreement between results obtained from multiple scanning iterations performed under identical conditions.[13,14]
Each Facial scanner has its own drawbacks: stereophotogrammetry-based scanners face challenges in precisely capturing hair and reflective or glossy skin surfaces, which limits its use in applications such as maxillofacial prosthetics.[15] Structured light scanners, on the other hand, use trigonometric triangulation to capture light patterns, forming a 3D surface model of the face.[16] These scanners typically use blue light, as it emits shorter wavelengths that are less prone to reflection, allowing for more precise scans compared to white light scanners, which are now being phased out. Light is captured from multiple angles, and a 3D mesh is generated based on the displacement of the light patterns. Laser scanning, which works similarly by capturing the reflection of a laser from the surface of the object, is highly accurate, though its precision can be compromised by external light sources. With the rapid advancement of 3D scanning technologies, determining the accuracy and reproducibility of scanned models has become crucial for their reliable use in clinical diagnostics. Each scanner has certain strengths and weaknesses across different facial regions, meaning a particular scanner may yield more accurate results in certain areas. Therefore, it is vital to verify these points and merge data from multiple scanners to produce the most accurate representation of soft tissue. While CBCT is established for its accuracy, there is limited data on how various facial scanners perform across different anatomical regions of the face, particularly for applications such as smile design and full-mouth rehabilitation. Our study aims to bridge this gap by systematically comparing the accuracy of three distinct facial scanners, identifying which device performs best in specific facial regions, and evaluating their potential as suitable replacements for CBCT in clinical workflows. The aim of this study is to analyze multiple facial scanners and check which scanner has the highest accuracy with the null hypothesis being that there is no difference between different scanners.
MATERIALS AND METHODS
Study design and sample size estimation
It is an in vivo study which includes four groups: Carestream face scan (Group A), Medit intraoral scanner (Group B), MetiSmile scanner (Group C), and Carestream CBCT (Group D). The study was approved by the university’s scientific ethical clearance committee (IHEC/SDC/PROSTHO-2203/23/204).
The sample size was determined using G*Power 3.1.9.3 for Mac OS X®(Heinrich-Heine-Universität Düsseldorf, Germany). A one-way analysis of variance (ANOVA) test was selected, with an effect size (Cohen’s f) 1.337 based on prior studies, a power of 0.80, and an alpha level of 0. 05. Eligible patients (n = 12) were consecutively enrolled from the university hospital’s outpatient clinic and randomized to scanning order using computer-generated numbers The sample size calculation was done using G Power (Version 3.0.10) based on an existing study. Level of significance was kept at 0.05 and power of study at 0.90.[17] The study included 36 scans (12 patients × 3 scanners), and using the calculated Cohen’s f = 1.337, we performed a post hoc power analysis. Post hoc power analysis confirmed 90% power for the observed effect size, aligning with recommendations for validating sample size adequacy.
Patient selection
The participants in the study were randomly chosen based on the patients who visited dental college and had gone for facial scanning and CBCT for orthodontic or full mouth rehab purpose; female patients or male patients having clean shave were selected for the study because facial hair coils hamper outcomes. Furthermore, patients with any form of scars or deformities were not selected to maintain uniformity in patient selection. Age was restricted between 18 and 30 years to maintain uniformity and homogenous population; patients were explained about the study and a written consent form was obtained. The exclusion criteria for this facial scanning study included patients under 18 or over 30 years of age, as well as those with significant facial trauma or deformities that could distort scanning results. Individuals unable to maintain a stable head position due to conditions such as severe neck or back disorders were also excluded. Patients with medical contraindications to CBCT, including pregnancy, were not eligible. In addition, individuals with active infections or open wounds on the facial skin, Facial hair, piercings, tattoos, scars, or implants that could obstruct or distort the scan were excluded. Patients with mental or cognitive impairments that would prevent informed consent or compliance with the scanning protocol were also excluded. Informed consent was obtained from all participants in accordance with the guidelines and approval of our institution’s human ethical research clearance board.
Patient preparation
Before scanning, patients were instructed to remove facial accessories like glasses and earrings and to tie back any hair covering their forehead. Male participants were asked to shave to minimize scanning inaccuracies, whereas female patients were asked to remove any cosmetic products on their face, to minimize reflection and inaccuracies. During the scan, patients were guided to keep their eyes and mouth gently closed and to refrain from making facial expressions or movements. The scanning environment, including lighting, was kept constant with a standardized lux of light during all scans.
Scanning protocol
Various surface scanning techniques were utilized to capture patient scans. These methods produced STL files as the output. To compare the surface scans, a CBCT was taken (Group D) for the same patient; this was obtained as a Digital Imaging and Communications in Medicine (DICOM) FILE. This DICOM file was later converted to STL to be able to compare to the test groups. No filtering or smoothing was applied to the STL files before analysis in order to check the accuracy of raw data. All the types of facial scans were taken for each patient included in the study [Figure 1]. The image capturing process was standardized, with each scan taking 20–30 s apart from the Medit scanner, which took approximately 3–4 min while ensuring a natural head position and controlled lighting.[18] Scanners were positioned at 30–40 cm from the subject for the facial scanner and 1 cm for Medit scanner, and all scans were performed under the observation of two trained operators. Inter-observer reliability was assessed using Intraclass Correlation Coefficient (ICC = 0.92, 95% confidence interval [CI]: 0.88–0.95), indicating excellent consistency.
Figure 1.
Three-dimensional (3D) facial scan captured by (a) Carestream CS9600 facial scanner; (b) Medit i700 intraoral scanner; (c) MetiSmile 3D face scanner
Group A (Carestream face scan)
Group A employed the Carestream (CS-9600) facial scanner, a CBCT machine integrated with a facial scanner mounted on a gantry. Operating on the principle of stereophotogrammetry, the scanner captures multiple images of the subject and combines them to generate 3D data. A notable advantage of this system is its ability to perform facial scans independently, without exposing the patient to ionizing radiation. During the scan, the subject was instructed to place their head on the designated headrest while the operator took the scan. The rotating gantry completed a single rotation around the patient to capture the scan. The rotation of the gantry covers areas from one ear to the ear and does not include the back of the head. The scan takes approximately less than a minute to complete.
Group B (Medit intra oral scanner)
For Group B, an intraoral scanner (Medit-i 700) was used to capture an extraoral scan. The Medit intraoral scanner is a handheld device that operates on structured light scanning, using a projected light pattern to capture 3D data. Equipped with advanced cameras and software, it generates high-resolution 3D images of the subject. The scanner’s software features a capability that allows for the scanning of extraoral soft tissues as well. Stage management option was selected in the Medit link app, and face scan was added, after which the scan was taken conventionally. The patient was positioned in a dental chair for the scan, which was conducted in a systematic order. Scanning began with the lips, followed by the chin, nose, left and right cheeks, eyes, and concluded with the forehead. This scanning approximately took 3–4 min per scan.
Group C (MetiSmile scanner)
Group C utilized the (MetiSmile) facial scanner, launched by Shining 3D in 2023. This dedicated facial scanner functions as both a handheld and tabletop device and operates using structured light scanning technology. For the scan, the patient was positioned against a neutral white background, and the operator systematically moved the scanner from left to right across the patient’s face to capture the required data. Each scan took <30 s.
Group D (Carestream cone beam computed tomography)
The control group was a conventional CBCT scan. This was taken using Carestream CS9600. The output was generated in DICOM format, which was later converted to STL for the purpose of this study. According to the literature, CBCT is considered to be the most accurate data to the real-time facial data of the patient.[19] Hence, CBCT data were chosen as the control group. The subjects were positioned with their Frankfort plane aligned parallel to the horizontal plane. They were instructed to gently close their eyes and mouth while avoiding any facial expressions. The patient was asked to occlude their teeth and rest their chin on a sponge, which was positioned on the chin rest, while the operator took the scan. CBCT Scan data can also be referred to as volumetric data as it takes multiple 2D X-rays from different angles and later reconstructs them as a 3D volume of the scanned surface. The exposure parameters were as follows: field of view – 160 mm × 170 mm, voxel size 150 microns, voltage 120 kVP, current 5 mA, and time 24 s.
Superimposition of data
A digitizing software (Radiant Dicom Viewer Version 4.2, Poland) was used to extract the outer skin surface from the CBCT [Figure 2] and convert it into STL. The validation was performed by an expert radiologist to check for potential errors. No filtering or smoothening was applied to the obtained data, and only the raw data were processed to ensure accuracy. The files were then imported into Geomagic Control X by 3D Systems (Rock Hill, SC, USA), an advanced software platform for digital inspection. The STL Data obtained from Radiant were used as a control group. This is referred to as measured data (MD) in Geomagic software. All the experimental groups (A, B, and C) were set as reference data (RD). To align these datasets, the software utilized its “initial alignment” feature, which facilitated a preliminary positional match. This was followed by the application of the “best-fit alignment” function, ensuring precise superimposition of the RD and MD images.
Figure 2.
Soft tissue volumetric data obtained from cone beam computed tomography
A difference of RD and MD images was measured at 12 reference points, namely Glabella (point 1), l right and left endocanthion (points 2 and 3), pronasale (point 4), left and right alare (points 5 and 6), subnasale (point 7), left and right Glogau–Klein point (gk point) (points 8 and 9), left and right chelion (points 10 and 11), and pogonion (point 12) [Figure 3]. These points were placed manually, and an ICC value for interobserver reliability was calculated. The ICC was 0.89 (95% CI: 0.82–0.94), indicating good to excellent reliability, ensuring consistency in reference point placement.[20] The software provided a comparison point tool to measure the deviation between the two data sets. The root mean square (RMS) value of deviation was recorded [Figure 4]. The range of color +1 mm (red) to −1 mm (blue) denotes the deviation, which was considered as clinically acceptable. A deviation in the range of +0.5 mm to −0.5 mm was considered to be preferable.[17]
Figure 3.
Illustration of 12 reference points, namely the Glabella (point 1), left and right right and left endocanthion (points 2 and 3), Pronasale (point 4), left and right alare (points 5 and 6), subnasale (Point 7), left and right Glogau-Klein point (gk point) (points 8 and 9), left and right chelion (points 10 and 11), and pogonion (point 12)
Figure 4.
Pointwise deviation of root mean square value from superimposition data using Geomagic
Each point was verified for accuracy by two observers. ICC values for interobserver reliability were calculated. The ICC was 0.89 (95% CI: 0.82–0.94), indicating good to excellent reliability, ensuring consistency in reference point placement.
Statistical test
Statistical tests were done using Statistical Package for Social Sciences Software (IBM Corp. Released 2011. IBM SPSS Statistics for Windows, Version 20.0. Armonk, NY: IBM Corp). The normality of the data was assessed using the Shapiro–Wilk test, which confirmed that the data followed a normal distribution (P > 0.05). Thus, parametric tests such as one-way ANOVA and post hoc Tukey tests were performed. One-way ANOVA was done to compare all three facial scans, and post hoc Tukey test was done for pairwise comparison.
RESULTS
All scanners demonstrated mean values below 1 mm, indicating clinically acceptable accuracy.[17] Notably, the mean values for Scanner B and Scanner C satisfied the “preferable” accuracy criterion of ± 0.5 mm. A one-way ANOVA analysis revealed a P = 0.001, indicating statistically significant differences of accuracy among the three groups. The mean deviation of scanner C was found to be the least (0.35 ± 0.33), indicating the highest overall accuracy amongst the three groups [Table 1]. Pairwise comparison for RMS deviation between groups showed significant difference was found when compared with Carestream Facial Scanner P = 0.001 and no significant difference was noted between MetiSmile and Medit scan [Table 2]. The eta-squared (η²) value for the ANOVA test is 0.641. This indicates a large effect size, meaning that the choice of facial scanner explains about 64.1% of the variance in RMS deviation.
Table 1.
Overall comparison among facial scanner groups (Carestream facial scanner, Medit, and MetiSmile) based on deviation of root mean square value
| Group | Mean±SD | 95% CI lower | 95% CI upper | SE | F | P |
|---|---|---|---|---|---|---|
| Carestream facial scanner (Group A) | 0.86±0.58 | 0.77 | 0.96 | 0.48 | 29.49 | 0.001* |
| Medit (Group B) | 0.48±0.78 | 0.35 | 0.61 | 0.65 | ||
| MetiSmile (Group C) | 0.35±0.33 | 0.29 | 0.40 | 0.27 |
*Significant at 0.05. P value was derived from a one-way ANOVA test. SD: Standard deviation, CI: Confidence interval, SE: Standard error
Table 2.
Pairwise comparisons between facial scanner (Carestream facial scanner, Medit, and MetiSmile) groups based on deviation
| Facial scan | MD | 95% CI lower | 95% CI upper | SE | P |
|---|---|---|---|---|---|
| Carestream facial scanner (Group A) versus MetiSmile (Group C) | 0.51 | 0.35 | 0.68 | 0.70 | 0.01* |
| Carestream facial scanner (Group A) versus Medit (Group B) | 0.38 | 0.22 | 0.55 | 0.70 | 0.01* |
| Medit (Group B) versus MetiSmile (Group C) | 0.13 | 0.29 | 0.34 | 0.70 | 0.10 |
*Significant at 0.05. P value was derived from post hoc Tukey test. CI: Confidence interval, SE: Standard error, MD: Measured data
The results correlated with the designated facial region to each point revealed that Scanner A performed better at the right ala of the nose, midpoint of the lower vermillion of the lip, and the left commissure of the lips. Scanner B showed superior performance at the glabella left and right endocanthion, philtrum, tip of cupid’s bow on the right side (Glogau–Klein [GK] point), right cheilion, and pogonion. Scanner C excelled in the glabella, nasion, left GK point, left and right endocanthion, right alare, pronasale, and right cheilion regions. Overall, Scanner C stood out to be the best scanner in terms of the least mean deviation and having maximum number of points with preferred values [Table 3].
Table 3.
Comparison among facial scanner groups (Carestream facial scanner, Medit, and MetiSmile) based on root mean square deviation on reference points
| Reference points | Facial scanner | Mean±SD | 95% CI lower | 95% CI upper | P |
|---|---|---|---|---|---|
| 1. Glabella | Carestream facial scanner (Group A) | 0.80±0.01 | 0.60 | 1.00 | 0.001* |
| Medit (Group B) | 0.97±0.01 | 0.77 | 1.10 | ||
| MetiSmile (Group C) | 0.23±0.01 | 0.20 | 0.40 | ||
| 2. Left endocanthion | Carestream facial scanner (Group A) | 1.326±0.02 | 1.10 | 1.50 | 0.001* |
| Medit (Group B) | 0.25±0.03 | 0.10 | 0.40 | ||
| MetiSmile (Group C) | 0.61±0.02 | 0.40 | 0.80 | ||
| 3. Right-endocanthion | Carestream facial scanner (Group A) | 1.19±0.01 | 0.90 | 1.30 | 0.001* |
| Medit (Group B) | 0.72±0.04 | 0.50 | 0.90 | ||
| MetiSmile (Group C) | 0.49±0.06 | 0.20 | 0.60 | ||
| 4. Pronasale | Carestream facial scanner (Group A) | 0.40±0.04 | 0.20 | 0.60 | 0.001* |
| Medit (Group B) | 0.20±0.03 | 0.10 | 0.40 | ||
| MetiSmile (Group C) | 1.35±0.05 | 1.20 | 1.47 | ||
| 5. Left alare | Carestream facial scanner (Group A) | 2.3±0.07 | 0.20 | 2.50 | 0.001* |
| Medit (Group B) | 0.81±0.02 | 0.60 | 1.01 | ||
| MetiSmile (Group C) | 0.15±0.01 | 0.12 | 0.17 | ||
| 6. Right - alare | Carestream facial scanner (Group-A) | 0.81±0.02 | 0.70 | 1.10 | 0.001* |
| Medit (Group B) | 0.3±0.03 | 0.10 | 0.52 | ||
| MetiSmile (Group C) | 0.15±0.03 | 0.12 | 0.37 | ||
| 7. Subnasale | Carestream facial scanner (Group A) | 0.94±0.04 | 0.70 | 1.10 | 0.001* |
| Medit (Group B) | 0.33±0.06 | 0.10 | 0.50 | ||
| MetiSmile (Group C) | 0.50±0.01 | 0.30 | 0.70 | ||
| 8. Right GK point | Carestream facial scanner (Group A) | 0.96±0.01 | 0.70 | 1.10 | 0.001* |
| Medit (Group B) | 0.15±0.04 | 0.12 | 0.35 | ||
| MetiSmile (Group C) | 0.24±0.03 | 0.20 | 0.40 | ||
| 9. Left GK point | Carestream facial scanner (Group A) | 0.22±0.03 | 0.20 | 0.40 | 0.001* |
| Medit (Group B) | 2.89±0.06 | 2.60 | 3.10 | ||
| MetiSmile (Group C) | 0.23±0.02 | 0.10 | 0.45 | ||
| 10. Left - cheilion | Carestream facial scanner (Group A) | 0.91±0.02 | 0.70 | 1.10 | 0.001* |
| Medit (Group B) | 0.31±0.07 | 0.10 | 0.50 | ||
| MetiSmile (Group C) | 0.17±0.04 | 0.11 | 0.30 | ||
| 11. Right - cheilion | Carestream facial scanner (Group A) | 0.63±0.09 | 0.40 | 0.80 | 0.001* |
| Medit (Group B) | 0.06±0.02 | 0.04 | 0.08 | ||
| MetiSmile (Group C) | 0.11±0.03 | 0.09 | 0.33 | ||
| 12. Pogonion | Carestream facial scanner (Group A) | 0.48±0.03 | 0.21 | 0.63 | 0.001* |
| Medit (Group B) | 1.6±0.04 | 1.50 | 1.80 | ||
| MetiSmile (Group C) | 0.22±0.05 | 0.11 | 0.24 |
*Significant at 0.05. P value was derived from a one-way ANOVA test. SD: Standard deviation, CI: Confidence interval, GK: Glogau–Klein
The results indicate that while all three facial scanners demonstrated clinically acceptable accuracy, the MetiSmile scanner exhibited the highest precision with the least mean deviation (0.35 ± 0.33 mm). This suggests that structured light scanning, as used in MetiSmile, provides superior accuracy compared to stereophotogrammetry-based and intraoral scanners adapted for facial scanning. Clinically, this has significant implications in fields such as maxillofacial prosthodontics, orthodontics, and digital smile design, where precise facial data is crucial for treatment planning and outcome predictability. The study supports the integration of dedicated facial scanners like MetiSmile into routine dental workflows, offering a noninvasive, highly accurate alternative to CBCT while reducing exposure to ionizing radiation.
DISCUSSION
MetiSmile utilizes structured light scanning technology, which has demonstrated superiority over the stereophotogrammetry used by the Carestream facial scanner.[21] This technological advantage likely explains MetiSmile’s superior performance in this study. When compared to the Medit scanner, which also employs structured light scanning, MetiSmile outperformed due to its higher sensor count, enabling more comprehensive facial capture. In addition, MetiSmile’s algorithm is specifically optimized for extraoral scanning, unlike Medit’s, which is adapted from its primary design as an intraoral scanner. This optimization contributes significantly to MetiSmile’s superior accuracy.[22]
The individual point results revealed that the Carestream facial scanner (Group A) excelled in capturing midface regions, such as the right ala of the nose, midpoint of the lower lip, and the left commissure of the lips. This accuracy is likely due to the camera’s angulation, optimized for midface scanning. Scanner B (Medit i700) performed best in the upper face, capturing points like the glabella, endocanthions, philtrum, cupid’s bow, and pogonion. However, its results were scattered, possibly due to operator skill and the scanner’s smaller head relative to the subject’s face. Scanner C (MetiSmile) demonstrated the most preferable points, including the glabella, GK point, endocanthions, right ala, pronasale, and right cheilion, attributed to its advanced surface light scanning technology and optimized algorithm.
Scanner A (Carestream CS9600) is a stationary device designed for fixed installation. Operators assist subjects in aligning their positions using visual markers for the ears and nose displayed on a monitor before capturing the image. Its 0.5-s capture time minimizes errors caused by patient movement, enhancing convenience for operators and subjects. However, its installation cost is a disadvantage, as not all CBCT machines include a facial scanner. Scanner A often showed distortions in curved regions, with recessed areas like the endocanthions appearing deeper compared to the reference image, this along with involuntary movement of the subjects or complex geometry of the face which the scanner was not able to capture as it rotated around a fixed axis Scanner B (Medit i700) primarily an intraoral scanner with an extraoral scanning feature, performed well in the upper face but was less accurate near the lips and chin. Its proximity to patients may lead to slight movements, especially in the lip region. In addition, its smaller scanning area requires multiple captures, resulting in interruptions and longer scanning times. Despite these drawbacks, its dual functionality as an intraoral and extraoral scanner provides significant cost savings, eliminating the need for a separate scanner.
Scanner C (MetiSmile), launched in 2023, is a dedicated extraoral facial scanner that proved the most accurate in this study, achieving preferable values in 8 out of 12 reference points, with two others in the clinically acceptable range. Its portability and scanning process, completed in under a minute, enhance its efficiency. The scanner’s algorithm allows operators to oversee the process in real time, adjust positioning, and address image gaps seamlessly. However, the operator’s skill heavily influences outcomes, and the flexibility of the device can sometimes make scanning cumbersome. Scanner C also exaggerated curvature in certain areas but showed smaller surface discrepancies and fewer deviations, highlighting its precision. It captured the undersides of protruding features, such as the subnasal and submandibular regions, more effectively than Scanner A due to its ability to capture from all angles.
Facial scanning technology in dentistry has advanced significantly with the rise of digital workflows. Traditional photographic protocols provide only 2D perspectives, which can lead to inaccuracies when aligning stereolithography (STL) files of digital wax-ups. Facial scanning addresses this limitation by capturing 3D scans, aligning teeth with facial planes while maintaining the proportional balance of facial thirds.[23,24] This technology enables the creation of a “virtual patient,” improving communication, treatment planning, and outcomes in prosthetic smile design, implant placement, and orthodontic treatments.[17,25] Creating a 3D image of the face relies on specialized sensors, categorized as active or passive. Active scanners emit signals that reflect off objects, while passive scanners use variations in infrared or visible light to construct images. Despite advancements, no unified scanning methodology exists, making studies like this critical. Such studies identify strengths and weaknesses of various scanners and provide insights for users and manufacturers to make informed choices. They also pave the way for developing technologies that integrate the best features of different scanners to create more accurate and reliable devices.
Mannequins could offer a controlled alternative for such studies, eliminating variables like subject movement and environmental inconsistencies. However, mannequins cannot replicate the diversity and complexity of human facial structures, limiting their applicability in real-world scenarios. A potential solution lies in integrating the strengths of all three scanners to develop a new technology or scanner that overcomes limitations such as operator-dependent errors and variability in facial expressions. This could lead to a unified scanning methodology capable of creating a flawless digital twin of the patient. The difference between different scanners will give the future scope where manufacturers can overcome all lacunae and make the technology better. The Medit i700 (₹7–8 lakhs) is an accurate intraoral scanner for restorative and orthodontic work.[26] The MetiSmile 3D (₹4.1–4.2 lakhs) enhances treatment planning with facial scans. The Carestream CS 9600 (₹60–70 lakhs) suits high-volume clinics with advanced diagnostic needs. Selection depends on the clinic needs and budget.
CONCLUSION
MetiSmile’s structured light scanning technology and optimized algorithm make it the most accurate scanner in this study, excelling in capturing a wide range of facial points. Carestream’s midface accuracy and Medit’s upper face performance highlight their respective strengths, but each scanner has limitations. The integration of their best features could lead to the development of a superior scanning technology, addressing current challenges and advancing digital workflows in dentistry.
Conflicts of interest
There are no conflicts of interest.
Funding Statement
Nil.
REFERENCES
- 1.Blasi A, Nucera R, Ronsivalle V, Candida E, Grippaudo C. Asymmetry index for the photogrammetric assessment of facial asymmetry. Am J Orthod Dentofacial Orthop. 2022;162:394–402. doi: 10.1016/j.ajodo.2021.04.030. [DOI] [PubMed] [Google Scholar]
- 2.Jain RK, Sowmithradevi S, Shantha SK. Correlation of forehead type with maxillary incisor inclination in Dravidian south Indian population: A prospective study. World J Dent. 2022;13:606–10. [Google Scholar]
- 3.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. doi: 10.1371/journal.pone.0169402. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Djordjevic J, Pirttiniemi P, Harila V, Heikkinen T, Toma AM, Zhurov AI, et al. Three-dimensional longitudinal assessment of facial symmetry in adolescents. Eur J Orthod. 2013;35:143–51. doi: 10.1093/ejo/cjr006. [DOI] [PubMed] [Google Scholar]
- 5.Al-Hiyali A, Ayoub A, Ju X, Almuzian M, Al-Anezi T. The impact of orthognathic surgery on facial expressions. J Oral Maxillofac Surg. 2015;73:2380–90. doi: 10.1016/j.joms.2015.05.008. [DOI] [PubMed] [Google Scholar]
- 6.Zaruba M, Ender A, Mehl A. New applications for three-dimensional follow-up and quality control using optical impression systems and OraCheck. Int J Comput Dent. 2014;17:53–64. [PubMed] [Google Scholar]
- 7.Toma AM, Zhurov A, Playle R, Ong E, Richmond S. Reproducibility of facial soft tissue landmarks on 3D laser-scanned facial images. Orthod Craniofac Res. 2009;12:33–42. doi: 10.1111/j.1601-6343.2008.01435.x. [DOI] [PubMed] [Google Scholar]
- 8.Bland JM, Altman DG. Agreed statistics: Measurement method comparison. Anesthesiology. 2012;116:182–5. doi: 10.1097/ALN.0b013e31823d7784. [DOI] [PubMed] [Google Scholar]
- 9.Agarwal S, Maiti S, Subhashree R. Acceptance towards smile makeover based on spa factor–A myth or reality. Int J Res Pharm Sci. 2020;11:1227–32. [Google Scholar]
- 10.Knoops PG, Beaumont CA, Borghi A, Rodriguez-Florez N, Breakey RW, Rodgers W, et al. Comparison of three-dimensional scanner systems for craniomaxillofacial imaging. J Plast Reconstr Aesthet Surg. 2017;70:441–9. doi: 10.1016/j.bjps.2016.12.015. [DOI] [PubMed] [Google Scholar]
- 11.Meyer-Marcotty P, Alpers GW, Gerdes AB, Stellzig-Eisenhauer A. Impact of facial asymmetry in visual perception: A 3-dimensional data analysis. Am J Orthod Dentofacial Orthop. 2010;137:168.e1–8. doi: 10.1016/j.ajodo.2008.11.023. [DOI] [PubMed] [Google Scholar]
- 12.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. doi: 10.1016/j.prosdent.2019.10.010. [DOI] [PubMed] [Google Scholar]
- 13.Bohner L, Gamba DD, Hanisch M, Marcio BS, Tortamano Neto P, Laganá DC, et al. Accuracy of digital technologies for the scanning of facial, skeletal, and intraoral tissues: A systematic review. J Prosthet Dent. 2019;121:246–51. doi: 10.1016/j.prosdent.2018.01.015. [DOI] [PubMed] [Google Scholar]
- 14.Grant GT, Campbell SD, Masri RM, Andersen MR, American College of Prosthodontists Digital Dentistry Glossary Development Task Force Glossary of digital dental terms: American College of Prosthodontists. J Prosthodont. 2016;25(Suppl 2):S2–9. doi: 10.1111/jopr.12532. [DOI] [PubMed] [Google Scholar]
- 15.Wadhwani V, Sivaswamy V, Rajaraman V. Surface roughness and marginal adaptation of stereolithography versus digital light processing three-dimensional printed resins: An in-vitro study. J Indian Prosthodont Soc. 2022;22:377–81. doi: 10.4103/jips.jips_8_22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sindhu JS, Maiti S, Nallaswamy D. Comparative analysis on efficiency and accuracy of parallel confocal microscopy and three-dimensional in motion video with triangulation technology-based intraoral scanner under influence of moisture and mouth opening – A crossover clinical trial. J Indian Prosthodont Soc. 2023;23:234–43. doi: 10.4103/jips.jips_65_23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cho RY, Byun SH, Yi SM, Ahn HJ, Nam YS, Park IY, et al. Comparative analysis of three facial scanners for creating digital twins by focusing on the difference in scanning method. Bioengineering (Basel) 2023;10:545. doi: 10.3390/bioengineering10050545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Maiti S, Sindhu S, Nallaswamy D. Accuracy and efficiency of two commercially available intraoral scanners under different room lighting conditions: A crossover clinical trial. Int J Prosthodont Restor Dent. 2023;13:201–9. [Google Scholar]
- 19.Hasegawa A, Shinya A, Lassila LV, Yokoyama D, Nakasone Y, Vallittu PK, et al. Accuracy of three-dimensional finite element modeling using two different dental cone beam computed tomography systems. Odontology. 2013;101:210–5. doi: 10.1007/s10266-012-0076-z. [DOI] [PubMed] [Google Scholar]
- 20.Di Blasio M, Minervini G, Segù M, Pedrazzi G, Di Blasio A, Cassi D, et al. An in-vivo study study on the effect of head orientation in the measurement of anthropometric points in stereophotogrammetry. Minerva Dent Oral Sci. 2024;73:343–51. doi: 10.23736/S2724-6329.24.04852-6. [DOI] [PubMed] [Google Scholar]
- 21.D’Ettorre G, Farronato M, Candida E, Quinzi V, Grippaudo C. A comparison between stereophotogrammetry and smartphone structured light technology for three-dimensional face scanning. Angle Orthod. 2022;92:358–63. doi: 10.2319/040921-290.1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Major M, Mészáros B, Würsching T, Polyák M, Kammerhofer G, Németh Z, et al. Evaluation of a structured light scanner for 3D facial imaging: A comparative study with direct anthropometry. Sensors (Basel) 2024;24:5286. doi: 10.3390/s24165286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Lee JD, Nguyen O, Lin YC, Luu D, Kim S, Amini A, et al. Facial scanners in dentistry: An overview. Prosthesis. 2022;4:664–78. [Google Scholar]
- 24.Jouhar R, Ahmed N, Ahmed MA, Faheemuddin M, Mosaddad SA, Heboyan A. Smile aesthetics in Pakistani population: Dentist preferences and perceptions of anterior teeth proportion and harmony. BMC Oral Health. 2024;24:401. doi: 10.1186/s12903-024-04100-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Rao DC, Vundavalli S, Indiran MA, Rao AK, Radhika D, Salloum MG. Role of oral health literacy in demand for oral healthcare services for missing teeth replacement among dental patients visiting a dental teaching hospital in India. J Indian Prosthodont Soc. 2025;25:59–66. doi: 10.4103/jips.jips_283_24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Alam MK, Alanazi NH, Alanzi TM, Alrwuili SM, Alazmi MS, Alruwaili AM, et al. Microesthetics in orthodontics: A systematic review and meta-analysis. J Orthod Sci. 2023;12:78. doi: 10.4103/jos.jos_84_23. [DOI] [PMC free article] [PubMed] [Google Scholar]