Three-Dimensional Imaging of the Face: A Comparison Between Three Different Imaging Modalities

Three-Dimensional Imaging of the Face: A Comparison Between Three Different Imaging Modalities Abstract Background Three-dimensional (3D) imaging of the face is being used extensively in medicine for clinical decision making, surgical planning, and research. Nowadays, several companies are offering a broad range of 3D imaging systems, varying in price, method, and mobility. However, most planning and evaluation methods are created and validated solely with one imaging system. Therefore, it is important to analyze possible differences in the 3D surface reconstruction between different systems. Objectives The objective of this study was to analyze differences in the 3D surface reconstruction between three systems: 3dMDface system, Vectra XT, and Artec Eva. Methods Three-dimensional images of the face were acquired from 15 healthy patients with each imaging system. Reproducibility of each device was calculated and a comparison of the Vectra XT and Artec Eva with the 3dMDface was made. Results All 3D imaging devices showed high reproducibility, with a mean difference of 0.18 ± 0.15 mm (3dMDface system), 0.15 ± 0.15 mm (Vectra XT), and 0.26 ± 0.24 mm (Artec Eva). No significant difference in reproducibility was found between the Vectra XT and 3dMDface, while a significant difference was found between 3dMDface and Artec Eva, and between Vectra XT and Artec Eva. The mean difference between 3dMDface and Vectra XT was 0.32 ± 0.26 mm. The mean difference between 3dMDface and Artec Eva was 0.44 ± 1.09 mm. Conclusions All three imaging devices showed high reproducibility and accuracy. Although the Artec Eva showed a significant lower reproducibility, the difference found was not clinically relevant. Therefore, using these different systems alongside each other in clinical and research settings is possible. Level of Evidence: 3 Three-dimensional stereophotogrammetry (3D imaging) is being used progressively in surgical specialisms for analysis, surgical planning, and research. Some examples are the evaluation of cleft lip surgery, 3D planning and follow up of facial surgery, 3D planning and evaluation of breast surgery, 3D evaluation of lymphedema, the evaluation of polygenetic populations, and for early screening of genetic anomalies.1-6 Besides the excellent and realistic representation of the captured body part, the advantages of 3D imaging are the absence of ionizing radiation, fast image capturing, and easy archival capabilities.7,8 Three-dimensional imaging systems are available widely and are able to capture 3D images within milliseconds. Due to this wide availability, as well as differing price ranges and possibilities, different systems are used even within the same hospital. Examples of systems commonly used in Dutch clinics are the 3dMD systems, the Vectra systems, and the Artec Eva. The 3dMD system is an active stereophotogrammetry system, of which the reliability and accuracy has been tested widely.7,9,10 The Vectra systems make use of the passive technique and also have been tested for accuracy and reliability.11,12 The Artec Eva is a handheld camera with known accuracy, which uses continues (active) scanning of the surface.13 The benefit of this system is that it is easy to transport and can be used in unconventional places and settings (eg, operating theatres). Although the separate 3D imaging systems have been validated for accuracy and reproducibility, no comparison between those particular systems is known. Since most research outcomes, analysis methods, and surgical planning techniques are based on a single imaging system, it is essential to know if these validated techniques can be applied when using a different imaging system. This would make comparison of research or the setup of multicenter studies with different imaging systems possible. Therefore, the goal of this study was to test the reproducibility of and to investigate the differences between three commonly used 3D imaging systems. METHODS Data Acquisition In December 2016, healthy patients who volunteered to participate in research were included in this study. Exclusion criteria were severe deformity of the face and the presence of facial hair. Three-dimensional images of the face were acquired by using three different imaging devices (Supplemental Figure 1, available online as Supplementary Material at www.aestheticsurgeryjournal.com). The first imaging setup was the 3dMDface system (3dMDface; 3dMD, Atlanta, GA), which consists of two pods with a total of six cameras. The second setup that was used was the Vectra XT (Canfield Imaging Systems, Canfield, OH), which has three pods with a total of six cameras. The last imaging setup was the Artec Eva (Artec, Luxembourg), which is a handheld scanning device that has three cameras. For the Artec Eva, the patient is scanned by moving the scanner manually around the patient. Three-dimensional images of the face were acquired twice with all imaging devices by an experienced user. For all 3D images, the patient was instructed to maintain a neutral facial expression during image acquisition. All patients gave written informed consent prior to data acquisition. The guidelines of the Declaration of Helsinki were followed during this study. Data Analysis For data analysis, all 3D images were loaded into Maxilim (Maxilim V2.3.0; Medicim, Mechelen, Belgium). To determine the reproducibility of every system (intrasystem accuracy), the first 3D image of every patient was matched onto the second 3D image for every imaging system separately (Figure 1). This was performed by a surface-based matching algorithm (iterative closest point [ICP]).14 To determine the differences between the two 3D images, from each point on the first image the closest distance towards the second image was calculated. The differences between the images were visualized by making a color-coded heat map (Figure 2). Figure 1. View largeDownload slide Schematic overview of the performed measurements. Figure 1. View largeDownload slide Schematic overview of the performed measurements. Figure 2. View largeDownload slide Overview of the data analysis. Two different images of a 24-year-old male (left) are matched onto each other with an ICP algorithm (middle). The differences between the two images are calculated and visualized with a color-coded heat map (right). Figure 2. View largeDownload slide Overview of the data analysis. Two different images of a 24-year-old male (left) are matched onto each other with an ICP algorithm (middle). The differences between the two images are calculated and visualized with a color-coded heat map (right). To determine the differences between the imaging systems (intersystem accuracy), the 3D images obtained by the Vectra XT and Artec Eva were matched onto the 3D images obtained by the 3dMDface system. To do this, the first-obtained 3D image for each imaging device was used for every patient. To determine the differences between both 3D images, the distance between both surfaces was calculated at each point. The region of the eyes is known to have significant errors in the 3dMD and Vectra 3D image; therefore, this region was excluded from the surface-based matching and further data analysis. Statistics To analyze the reproducibility and the differences between 3D imaging devices, the absolute mean distance, the 95th percentile (p95), and standard deviation (SD) were calculated for each matching of the 3D images. One-way ANOVA with post hoc analyses was performed to determine differences between imaging devices. One-sample t tests were performed to determine if the difference between the Vectra XT or Artec Eva and the 3dMDface system were different from zero. A P value of <0.05 was considered statistically significant. All statistics were performed using IBM SPSS Statistics, Version 22 (IBM Corp., Armonk, NY). RESULTS Fifteen healthy patients were included in this study (6 male and 9 female; mean age, 37 ± 12 years; age range, 24-59 years). Intrasystem Reproducibility The mean absolute difference, SD, and p95 between the two 3D images acquired with the same scanning system are shown in Table 1. One-way ANOVA analysis showed a significant difference between the imaging systems (sum of squares = 0.101, F = 5.499, P = 0.008). Post hoc multiple comparisons analysis showed a significant difference between the Artec Eva and the 3dMDface system (0.9 mm, P = 0.016), and between Artec Eva and Vectra XT (0.1 mm, P = 0.003). No significant difference was found between the 3dMDface system and the Vectra XT (P = 0.535). A box plot of the results is shown in Figure 3. Table 1. Reproducibility of Each Imaging System Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm View Large Table 1. Reproducibility of Each Imaging System Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm View Large Figure 3. View largeDownload slide Reproducibility of the different imaging systems. The circles represent statistical outliers for patients 1, 10, and 11. The asterisks represents an extreme outlier for patient 1. Figure 3. View largeDownload slide Reproducibility of the different imaging systems. The circles represent statistical outliers for patients 1, 10, and 11. The asterisks represents an extreme outlier for patient 1. Intersystem Reproducibility The mean absolute differences between the Vectra XT and 3dMDface system and between the Artec Eva and 3dMDface system are shown in Table 2. The mean absolute difference between the Vectra XT and 3dMDface system was 0.32 ± 0.26 mm. The mean absolute difference between the Artec Eva and 3dMDface system was 0.44 ± 1.09 mm. One-sample t tests showed a significant difference for both the Vectra XT (P < 0.00) and the Artec Eva (P < 0.00) compared with the 3dMD system. A box plot of the results is shown in Figure 4. A color-coded heat map of the differences between the systems is shown in Figure 5. Table 2. Difference Between 3dMD Facial and Vectra XT/Artec Eva Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm View Large Table 2. Difference Between 3dMD Facial and Vectra XT/Artec Eva Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm View Large Figure 4. View largeDownload slide Mean absolute difference between the Vectra XT and the 3dMDface system (left) and the mean absolute difference between the Artec Eva and the 3dMDface system (right). The circles represent statistical outliers for patients 1 and 13. Figure 4. View largeDownload slide Mean absolute difference between the Vectra XT and the 3dMDface system (left) and the mean absolute difference between the Artec Eva and the 3dMDface system (right). The circles represent statistical outliers for patients 1 and 13. Figure 5. View largeDownload slide A representative, color-coded heat map of the differences between (A) the 3dMDface system and Vectra XT, and between (B) the 3dMDface system and Artec Eva, based on the 3D images of a 25-year-old male. As can be seen, the resulting 3D image from the Artec Eva is overall bigger, while the Vectra XT shows a more random difference pattern. Figure 5. View largeDownload slide A representative, color-coded heat map of the differences between (A) the 3dMDface system and Vectra XT, and between (B) the 3dMDface system and Artec Eva, based on the 3D images of a 25-year-old male. As can be seen, the resulting 3D image from the Artec Eva is overall bigger, while the Vectra XT shows a more random difference pattern. DISCUSSION This study investigated the reproducibility and accuracy of three different 3D imaging systems. To the best of our knowledge, this is the first study that compares these three systems for facial imaging. The 3dMDface system and the Vectra XT showed the highest reproducibility (0.18 ± 0.15 mm and 0.15 ± 0.15 mm, respectively), which did not differ significantly from each other (P = 0.535). The Artec Eva showed a reproducibility of 0.26 ± 0.24 mm, which was significantly different from both the 3dMDface system and Vectra XT (P = 0.016 and P = 0.003, respectively). To evaluate the differences between the imaging systems, both the Vectra XT 3D image and the Artec Eva 3D image were matched onto the 3dMD 3D image. The 3dMD is the imaging system that is covered most extensively in the literature, and therefore this system was chosen as the golden standard.9,10,14-16 The Vectra XT showed a mean difference with the 3dMDface system of 0.32 ± 0.26, which was significantly different (P < 0.00). The Artec Eva showed a mean difference with the 3dMDface system of 0.44 ± 1.09 mm, which was also significantly different (P < 0.00). A representative heat map of the differences between the systems is shown in Figure 5. While the differences between the 3dMDface and Vectra XT give a random pattern, the 3D image made with the Artec Eva is overall bigger, resulting in an entire green difference map. Literature Dindaroğlu et al investigated the accuracy of the 3dMD system by performing anthropomorphic measurements on 80 patients, and reproducing those measurements on the 3D image made by the 3dMD system.7 They reported an average error of 0.21 mm. Weinberg et al performed a similar study to investigate anthropomorphic accuracy of the 3dMD system and reported an average accuracy of 0.26 mm by comparing anthropomorphic measurements on mannequin heads with measurements based on the 3dMD 3D image.10 De Menezes et al investigated the accuracy of the Vectra system, which was reported as 0.02 ± 0.03 mm by measuring a cubic object. The intrasystem reproducibility was investigated by twice measuring the distance between several landmarks on the face of 10 patients. They reported a mean absolute difference between 0.13 (nasion-subnasale distance) and 1.19 mm (mouth width).17 An accuracy of 1.33 mm was found by Metzler et al for anthropomorphic measurements on a mannequin head, after excluding 10% of the measured landmarks due to incorrect measurements.11 For the Artec Eva, Modabber et al reported a mean error of 0.23 ± 0.05 mm by measuring the error of a scanned Lego brick attached to the forehead of a patient.13 A mean error of 0.24 ± 0.09 was reported with the same method for the cheek region. These studies justify our choice to consider 3dMD as our golden standard. The reproducibility and accuracy that was found in our study is in the same range as the accuracy of these systems reported in literature. However, our measurement method is slightly different. All of these studies report accuracy by measuring distances between landmarks in the face. Therefore, they have only a limited number of measurement points in the face. Our study measures the difference between two 3D images based on all data points that are present in the 3D image. Therefore, we show the error of the total 3D image. On the other hand, the disadvantage of our method is that we can draw conclusions only for the whole face at once, and cannot distinguish errors in different parts of the face. No comparisons between the 3dMD face, Vectra XT, and Artec Eva were found in the literature. Possible Errors As the region of the eyes is known to have significant errors in the 3dMD and Vectra 3D image, this region was excluded from the surface-based matching and data analysis (Figure 6).18 Therefore, the results described in this article do not apply to this region. Figure 6. View largeDownload slide As the region of the eye shows significant errors in all 3D images from all systems, this region was excluded from the analysis. The example shows the 3D images of a 24-year-old male. Figure 6. View largeDownload slide As the region of the eye shows significant errors in all 3D images from all systems, this region was excluded from the analysis. The example shows the 3D images of a 24-year-old male. While the 3dMDface system and the Vectra XT both acquire several images at the same moment (acquisition time ~1.5 ms and ~3.5 ms, respectively), the Artec Eva uses continuous scanning of the face to construct a 3D image. Scanning of the face with the Artec Eva takes approximately 20 seconds for experienced users. During this relatively long scanning time, it is possible that minor differences in the facial expression of the patient appear, leading to a less accurate 3D image. This might explain the lower reproducibility and accuracy found for the Artec Eva. Advantages and Disadvantages An advantage of the Artec Eva is that it is a mobile scanner, giving the opportunity to make 3D images on any desired location. However, more experience and training is necessary to be able to correctly capture correct 3D images of the face. Furthermore, manual data processing is required before the 3D image is adequate for analysis or surgical planning, which will cost an experienced user a few minutes. Therefore, imaging with the Artec Eva is often performed by an experienced researcher. The 3dMDface system and Vectra XT are not mobile and have higher costs. However, those systems require less training, have faster image acquisition and capture the complete face in a single moment (1.5 ms and 3.5 ms, respectively). Acquiring images with these systems in a clinical workflow is fast and easy and often performed by a nurse in the outpatient clinic. With a price of less than €15,000, the Artec Eva is the cheapest of the systems. Both the 3dMDface system and Vectra XT cost around €40,000. The Vectra XT also offers the possibility to make 3D images of the torso, while the 3dMD system can be upgraded by adding more pods, resulting in a bigger field of view. Also, the 3dMD pods can be adjusted easily in settings and position, making the system more flexible and suitable for research purposes and specific clinical applications. Differences Between Systems Although the differences in the reconstructed surfaces found between the imaging systems are significant, they are not directly clinically relevant, as they were all under 0.5 mm. Furthermore, a small difference in the patient’s face will occur in between image acquisitions. This is known to be around 0.25 mm average for sequential images, and 0.36 mm for images with a time difference of six weeks.19 However, these differences will be present for every imaging system. This indicates that 3D images of those three systems are comparable with each other in terms of accuracy, and that these systems can be used simultaneously in, for example, multicenter studies. CONCLUSION The 3dMDface system and the Vectra XT have similar reproducibility, while the Artec Eva has significantly lower reproducibility. The Vectra XT and Artec Eva showed a small mean difference with the 3D camera from 3dMD. However, the differences found were not clinically relevant, indicating that 3D images of all three systems are comparable and can be used simultaneously in multicenter studies. The main advantage of the Artec Eva is that it is a mobile scanner, while the 3dMDface system and Vectra XT are rigid systems that cannot be transported easily. Also, the costs of the Artec Eva are significantly lower than the costs of the 3dMDface system and Vectra XT. However, the Artec Eva requires more acquisition time and is therefore more prone to movement artifacts. Furthermore, the data processing for the Artec Eva is performed manually. In contrast, the 3dMDface system and Vectra XT have fast acquisition times and have 3D images almost instantly available for analysis and surgical planning. Supplementary Material This article contains supplementary material located online at www.aestheticsurgeryjournal.com. Disclosures The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. REFERENCES 1. Meulstee J , Liebregts J , Xi T et al. A new 3D approach to evaluate facial profile changes following BSSO . J Craniomaxillofac Surg . 2015 ; 43 ( 10 ): 1994 - 1999 . Google Scholar CrossRef Search ADS PubMed 2. van Loon B , Maal TJ , Plooij JM et al. 3D Stereophotogrammetric assessment of pre- and postoperative volumetric changes in the cleft lip and palate nose . Int J Oral Maxillofac Surg . 2010 ; 39 ( 6 ): 534 - 540 . Google Scholar CrossRef Search ADS PubMed 3. Hopman SMJ , Merks JHM , Suttie M , Hennekam RCM , Hammond P . Face shape differs in phylogenetically related populations . Eur J Hum Genet . 2014 ; 22 ( 11 ): 1268 - 1271 . Google Scholar CrossRef Search ADS PubMed 4. Baan F , Liebregts J , Xi T et al. A New 3D Tool for Assessing the Accuracy of Bimaxillary Surgery: The OrthoGnathicAnalyser . PLoS One . 2016 ; 11 ( 2 ): e0149625 . Google Scholar CrossRef Search ADS PubMed 5. Hameeteman M , Verhulst AC , Maal TJJ , Ulrich DJO . An analysis of pose in 3D stereophotogrammetry of the breast . J Plast Reconstr Aesthetic Surg . 2016 ; 69 ( 12 ): 1609 - 1613 . Google Scholar CrossRef Search ADS 6. Hameeteman M , Verhulst AC , Vreeken RD , Maal TJJ , Ulrich DJO . 3D stereophotogrammetry in upper-extremity lymphedema: An accurate diagnostic method . J Plast Reconstr Aesthetic Surg . 2016 ; 69 ( 2 ): 241 - 247 . Google Scholar CrossRef Search ADS 7. 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 ( 3 ): 487 - 494 . Google Scholar CrossRef Search ADS PubMed 8. Plooij JM , Swennen GRJ , Rangel FA et al. Evaluation of reproducibility and reliability of 3D soft tissue analysis using 3D stereophotogrammetry . Int J Oral Maxillofac Surg . 2009 ; 38 ( 3 ): 267 - 273 . Google Scholar CrossRef Search ADS PubMed 9. Lübbers H-T , Medinger L , Kruse A , Grätz KW , Matthews F . Precision and Accuracy of the 3dMD Photogrammetric System in Craniomaxillofacial Application . J Craniofac Surg . 2010 ; 21 ( 3 ): 763 - 767 . Google Scholar CrossRef Search ADS PubMed 10. 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 ( 3 ): 477 - 483 . Google Scholar CrossRef Search ADS PubMed 11. Metzler P , Sun Y , Zemann W et al. Validity of the 3D VECTRA photogrammetric surface imaging system for cranio-maxillofacial anthropometric measurements . Oral Maxillofac Surg . 2014 ; 18 ( 3 ): 297 - 304 . Google Scholar CrossRef Search ADS PubMed 12. Spanholtz T , Leitsch S , Holzbach T , Volkmer E , Engelhardt T , Giunta R . 3-dimensionale Bilderfassung: Erste Erfahrungen in der Planung und Dokumentation plastisch-chirurgischer Operationen . Handchirurgie Mikrochirurgie Plast Chir . 2012 ; 44 ( 4 ): 234 - 239 . Google Scholar CrossRef Search ADS 13. Modabber A , Peters F , Kniha K et al. Evaluation of the accuracy of a mobile and a stationary system for three-dimensional facial scanning . J Craniomaxillofac Surg . 2016 ; 44 ( 10 ): 1719 - 1724 . Google Scholar CrossRef Search ADS PubMed 14. Maal TJJ , van Loon B , Plooij JM et al. Registration of 3-dimensional facial photographs for clinical use . J Oral Maxillofac Surg . 2010 ; 68 ( 10 ): 2391 - 2401 . Google Scholar CrossRef Search ADS PubMed 15. 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 ( 3 ): 247 - 253 . Google Scholar CrossRef Search ADS PubMed 16. Metzger TE , Kula KS , Eckert GJ , Ghoneima AA . Orthodontic soft-tissue parameters: a comparison of cone-beam computed tomography and the 3dMD imaging system . Am J Orthod Dentofacial Orthop . 2013 ; 144 ( 5 ): 672 - 681 . Google Scholar CrossRef Search ADS PubMed 17. De Menezes M , Rosati R , Ferrario VF , Sforza C . Accuracy and reproducibility of a 3-dimensional stereophotogrammetric imaging system . J Oral Maxillofac Surg . 2010 ; 68 ( 9 ): 2129 - 2135 . Google Scholar CrossRef Search ADS PubMed 18. Maal TJJ , Plooij JM , Rangel FA , Mollemans W , Schutyser FAC , Bergé SJ . The accuracy of matching three-dimensional photographs with skin surfaces derived from cone-beam computed tomography . Int J Oral Maxillofac Surg . 2008 ; 37 ( 7 ): 641 - 646 . Google Scholar CrossRef Search ADS PubMed 19. Maal TJJ , Verhamme LM , van Loon B et al. . Variation of the face in rest using 3D stereophotogrammetry . Int J Oral Maxillofac Surg . 2011 ; 40 : 1252 - 1257 . © 2018 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Aesthetic Surgery Journal Oxford University Press

Three-Dimensional Imaging of the Face: A Comparison Between Three Different Imaging Modalities

Loading next page...
 
/lp/ou_press/three-dimensional-imaging-of-the-face-a-comparison-between-three-zq1qTNpEX6
Publisher
Mosby Inc.
Copyright
© 2018 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com
ISSN
1090-820X
eISSN
1527-330X
D.O.I.
10.1093/asj/sjx227
Publisher site
See Article on Publisher Site

Abstract

Abstract Background Three-dimensional (3D) imaging of the face is being used extensively in medicine for clinical decision making, surgical planning, and research. Nowadays, several companies are offering a broad range of 3D imaging systems, varying in price, method, and mobility. However, most planning and evaluation methods are created and validated solely with one imaging system. Therefore, it is important to analyze possible differences in the 3D surface reconstruction between different systems. Objectives The objective of this study was to analyze differences in the 3D surface reconstruction between three systems: 3dMDface system, Vectra XT, and Artec Eva. Methods Three-dimensional images of the face were acquired from 15 healthy patients with each imaging system. Reproducibility of each device was calculated and a comparison of the Vectra XT and Artec Eva with the 3dMDface was made. Results All 3D imaging devices showed high reproducibility, with a mean difference of 0.18 ± 0.15 mm (3dMDface system), 0.15 ± 0.15 mm (Vectra XT), and 0.26 ± 0.24 mm (Artec Eva). No significant difference in reproducibility was found between the Vectra XT and 3dMDface, while a significant difference was found between 3dMDface and Artec Eva, and between Vectra XT and Artec Eva. The mean difference between 3dMDface and Vectra XT was 0.32 ± 0.26 mm. The mean difference between 3dMDface and Artec Eva was 0.44 ± 1.09 mm. Conclusions All three imaging devices showed high reproducibility and accuracy. Although the Artec Eva showed a significant lower reproducibility, the difference found was not clinically relevant. Therefore, using these different systems alongside each other in clinical and research settings is possible. Level of Evidence: 3 Three-dimensional stereophotogrammetry (3D imaging) is being used progressively in surgical specialisms for analysis, surgical planning, and research. Some examples are the evaluation of cleft lip surgery, 3D planning and follow up of facial surgery, 3D planning and evaluation of breast surgery, 3D evaluation of lymphedema, the evaluation of polygenetic populations, and for early screening of genetic anomalies.1-6 Besides the excellent and realistic representation of the captured body part, the advantages of 3D imaging are the absence of ionizing radiation, fast image capturing, and easy archival capabilities.7,8 Three-dimensional imaging systems are available widely and are able to capture 3D images within milliseconds. Due to this wide availability, as well as differing price ranges and possibilities, different systems are used even within the same hospital. Examples of systems commonly used in Dutch clinics are the 3dMD systems, the Vectra systems, and the Artec Eva. The 3dMD system is an active stereophotogrammetry system, of which the reliability and accuracy has been tested widely.7,9,10 The Vectra systems make use of the passive technique and also have been tested for accuracy and reliability.11,12 The Artec Eva is a handheld camera with known accuracy, which uses continues (active) scanning of the surface.13 The benefit of this system is that it is easy to transport and can be used in unconventional places and settings (eg, operating theatres). Although the separate 3D imaging systems have been validated for accuracy and reproducibility, no comparison between those particular systems is known. Since most research outcomes, analysis methods, and surgical planning techniques are based on a single imaging system, it is essential to know if these validated techniques can be applied when using a different imaging system. This would make comparison of research or the setup of multicenter studies with different imaging systems possible. Therefore, the goal of this study was to test the reproducibility of and to investigate the differences between three commonly used 3D imaging systems. METHODS Data Acquisition In December 2016, healthy patients who volunteered to participate in research were included in this study. Exclusion criteria were severe deformity of the face and the presence of facial hair. Three-dimensional images of the face were acquired by using three different imaging devices (Supplemental Figure 1, available online as Supplementary Material at www.aestheticsurgeryjournal.com). The first imaging setup was the 3dMDface system (3dMDface; 3dMD, Atlanta, GA), which consists of two pods with a total of six cameras. The second setup that was used was the Vectra XT (Canfield Imaging Systems, Canfield, OH), which has three pods with a total of six cameras. The last imaging setup was the Artec Eva (Artec, Luxembourg), which is a handheld scanning device that has three cameras. For the Artec Eva, the patient is scanned by moving the scanner manually around the patient. Three-dimensional images of the face were acquired twice with all imaging devices by an experienced user. For all 3D images, the patient was instructed to maintain a neutral facial expression during image acquisition. All patients gave written informed consent prior to data acquisition. The guidelines of the Declaration of Helsinki were followed during this study. Data Analysis For data analysis, all 3D images were loaded into Maxilim (Maxilim V2.3.0; Medicim, Mechelen, Belgium). To determine the reproducibility of every system (intrasystem accuracy), the first 3D image of every patient was matched onto the second 3D image for every imaging system separately (Figure 1). This was performed by a surface-based matching algorithm (iterative closest point [ICP]).14 To determine the differences between the two 3D images, from each point on the first image the closest distance towards the second image was calculated. The differences between the images were visualized by making a color-coded heat map (Figure 2). Figure 1. View largeDownload slide Schematic overview of the performed measurements. Figure 1. View largeDownload slide Schematic overview of the performed measurements. Figure 2. View largeDownload slide Overview of the data analysis. Two different images of a 24-year-old male (left) are matched onto each other with an ICP algorithm (middle). The differences between the two images are calculated and visualized with a color-coded heat map (right). Figure 2. View largeDownload slide Overview of the data analysis. Two different images of a 24-year-old male (left) are matched onto each other with an ICP algorithm (middle). The differences between the two images are calculated and visualized with a color-coded heat map (right). To determine the differences between the imaging systems (intersystem accuracy), the 3D images obtained by the Vectra XT and Artec Eva were matched onto the 3D images obtained by the 3dMDface system. To do this, the first-obtained 3D image for each imaging device was used for every patient. To determine the differences between both 3D images, the distance between both surfaces was calculated at each point. The region of the eyes is known to have significant errors in the 3dMD and Vectra 3D image; therefore, this region was excluded from the surface-based matching and further data analysis. Statistics To analyze the reproducibility and the differences between 3D imaging devices, the absolute mean distance, the 95th percentile (p95), and standard deviation (SD) were calculated for each matching of the 3D images. One-way ANOVA with post hoc analyses was performed to determine differences between imaging devices. One-sample t tests were performed to determine if the difference between the Vectra XT or Artec Eva and the 3dMDface system were different from zero. A P value of <0.05 was considered statistically significant. All statistics were performed using IBM SPSS Statistics, Version 22 (IBM Corp., Armonk, NY). RESULTS Fifteen healthy patients were included in this study (6 male and 9 female; mean age, 37 ± 12 years; age range, 24-59 years). Intrasystem Reproducibility The mean absolute difference, SD, and p95 between the two 3D images acquired with the same scanning system are shown in Table 1. One-way ANOVA analysis showed a significant difference between the imaging systems (sum of squares = 0.101, F = 5.499, P = 0.008). Post hoc multiple comparisons analysis showed a significant difference between the Artec Eva and the 3dMDface system (0.9 mm, P = 0.016), and between Artec Eva and Vectra XT (0.1 mm, P = 0.003). No significant difference was found between the 3dMDface system and the Vectra XT (P = 0.535). A box plot of the results is shown in Figure 3. Table 1. Reproducibility of Each Imaging System Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm View Large Table 1. Reproducibility of Each Imaging System Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm Mean absolute difference Standard deviation 95th percentile (P95) 3dMDface 0.18 mm 0.15 mm 0.47 mm Vectra XT 0.15 mm 0.15 mm 0.45 mm Artec Eva 0.26 mm 0.24 mm 0.74 mm View Large Figure 3. View largeDownload slide Reproducibility of the different imaging systems. The circles represent statistical outliers for patients 1, 10, and 11. The asterisks represents an extreme outlier for patient 1. Figure 3. View largeDownload slide Reproducibility of the different imaging systems. The circles represent statistical outliers for patients 1, 10, and 11. The asterisks represents an extreme outlier for patient 1. Intersystem Reproducibility The mean absolute differences between the Vectra XT and 3dMDface system and between the Artec Eva and 3dMDface system are shown in Table 2. The mean absolute difference between the Vectra XT and 3dMDface system was 0.32 ± 0.26 mm. The mean absolute difference between the Artec Eva and 3dMDface system was 0.44 ± 1.09 mm. One-sample t tests showed a significant difference for both the Vectra XT (P < 0.00) and the Artec Eva (P < 0.00) compared with the 3dMD system. A box plot of the results is shown in Figure 4. A color-coded heat map of the differences between the systems is shown in Figure 5. Table 2. Difference Between 3dMD Facial and Vectra XT/Artec Eva Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm View Large Table 2. Difference Between 3dMD Facial and Vectra XT/Artec Eva Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm Mean absolute difference Standard deviation Vectra XT vs 3dMDface 0.32 mm 0.26 mm Artec Eva vs 3dMDface 0.44 mm 1.09 mm View Large Figure 4. View largeDownload slide Mean absolute difference between the Vectra XT and the 3dMDface system (left) and the mean absolute difference between the Artec Eva and the 3dMDface system (right). The circles represent statistical outliers for patients 1 and 13. Figure 4. View largeDownload slide Mean absolute difference between the Vectra XT and the 3dMDface system (left) and the mean absolute difference between the Artec Eva and the 3dMDface system (right). The circles represent statistical outliers for patients 1 and 13. Figure 5. View largeDownload slide A representative, color-coded heat map of the differences between (A) the 3dMDface system and Vectra XT, and between (B) the 3dMDface system and Artec Eva, based on the 3D images of a 25-year-old male. As can be seen, the resulting 3D image from the Artec Eva is overall bigger, while the Vectra XT shows a more random difference pattern. Figure 5. View largeDownload slide A representative, color-coded heat map of the differences between (A) the 3dMDface system and Vectra XT, and between (B) the 3dMDface system and Artec Eva, based on the 3D images of a 25-year-old male. As can be seen, the resulting 3D image from the Artec Eva is overall bigger, while the Vectra XT shows a more random difference pattern. DISCUSSION This study investigated the reproducibility and accuracy of three different 3D imaging systems. To the best of our knowledge, this is the first study that compares these three systems for facial imaging. The 3dMDface system and the Vectra XT showed the highest reproducibility (0.18 ± 0.15 mm and 0.15 ± 0.15 mm, respectively), which did not differ significantly from each other (P = 0.535). The Artec Eva showed a reproducibility of 0.26 ± 0.24 mm, which was significantly different from both the 3dMDface system and Vectra XT (P = 0.016 and P = 0.003, respectively). To evaluate the differences between the imaging systems, both the Vectra XT 3D image and the Artec Eva 3D image were matched onto the 3dMD 3D image. The 3dMD is the imaging system that is covered most extensively in the literature, and therefore this system was chosen as the golden standard.9,10,14-16 The Vectra XT showed a mean difference with the 3dMDface system of 0.32 ± 0.26, which was significantly different (P < 0.00). The Artec Eva showed a mean difference with the 3dMDface system of 0.44 ± 1.09 mm, which was also significantly different (P < 0.00). A representative heat map of the differences between the systems is shown in Figure 5. While the differences between the 3dMDface and Vectra XT give a random pattern, the 3D image made with the Artec Eva is overall bigger, resulting in an entire green difference map. Literature Dindaroğlu et al investigated the accuracy of the 3dMD system by performing anthropomorphic measurements on 80 patients, and reproducing those measurements on the 3D image made by the 3dMD system.7 They reported an average error of 0.21 mm. Weinberg et al performed a similar study to investigate anthropomorphic accuracy of the 3dMD system and reported an average accuracy of 0.26 mm by comparing anthropomorphic measurements on mannequin heads with measurements based on the 3dMD 3D image.10 De Menezes et al investigated the accuracy of the Vectra system, which was reported as 0.02 ± 0.03 mm by measuring a cubic object. The intrasystem reproducibility was investigated by twice measuring the distance between several landmarks on the face of 10 patients. They reported a mean absolute difference between 0.13 (nasion-subnasale distance) and 1.19 mm (mouth width).17 An accuracy of 1.33 mm was found by Metzler et al for anthropomorphic measurements on a mannequin head, after excluding 10% of the measured landmarks due to incorrect measurements.11 For the Artec Eva, Modabber et al reported a mean error of 0.23 ± 0.05 mm by measuring the error of a scanned Lego brick attached to the forehead of a patient.13 A mean error of 0.24 ± 0.09 was reported with the same method for the cheek region. These studies justify our choice to consider 3dMD as our golden standard. The reproducibility and accuracy that was found in our study is in the same range as the accuracy of these systems reported in literature. However, our measurement method is slightly different. All of these studies report accuracy by measuring distances between landmarks in the face. Therefore, they have only a limited number of measurement points in the face. Our study measures the difference between two 3D images based on all data points that are present in the 3D image. Therefore, we show the error of the total 3D image. On the other hand, the disadvantage of our method is that we can draw conclusions only for the whole face at once, and cannot distinguish errors in different parts of the face. No comparisons between the 3dMD face, Vectra XT, and Artec Eva were found in the literature. Possible Errors As the region of the eyes is known to have significant errors in the 3dMD and Vectra 3D image, this region was excluded from the surface-based matching and data analysis (Figure 6).18 Therefore, the results described in this article do not apply to this region. Figure 6. View largeDownload slide As the region of the eye shows significant errors in all 3D images from all systems, this region was excluded from the analysis. The example shows the 3D images of a 24-year-old male. Figure 6. View largeDownload slide As the region of the eye shows significant errors in all 3D images from all systems, this region was excluded from the analysis. The example shows the 3D images of a 24-year-old male. While the 3dMDface system and the Vectra XT both acquire several images at the same moment (acquisition time ~1.5 ms and ~3.5 ms, respectively), the Artec Eva uses continuous scanning of the face to construct a 3D image. Scanning of the face with the Artec Eva takes approximately 20 seconds for experienced users. During this relatively long scanning time, it is possible that minor differences in the facial expression of the patient appear, leading to a less accurate 3D image. This might explain the lower reproducibility and accuracy found for the Artec Eva. Advantages and Disadvantages An advantage of the Artec Eva is that it is a mobile scanner, giving the opportunity to make 3D images on any desired location. However, more experience and training is necessary to be able to correctly capture correct 3D images of the face. Furthermore, manual data processing is required before the 3D image is adequate for analysis or surgical planning, which will cost an experienced user a few minutes. Therefore, imaging with the Artec Eva is often performed by an experienced researcher. The 3dMDface system and Vectra XT are not mobile and have higher costs. However, those systems require less training, have faster image acquisition and capture the complete face in a single moment (1.5 ms and 3.5 ms, respectively). Acquiring images with these systems in a clinical workflow is fast and easy and often performed by a nurse in the outpatient clinic. With a price of less than €15,000, the Artec Eva is the cheapest of the systems. Both the 3dMDface system and Vectra XT cost around €40,000. The Vectra XT also offers the possibility to make 3D images of the torso, while the 3dMD system can be upgraded by adding more pods, resulting in a bigger field of view. Also, the 3dMD pods can be adjusted easily in settings and position, making the system more flexible and suitable for research purposes and specific clinical applications. Differences Between Systems Although the differences in the reconstructed surfaces found between the imaging systems are significant, they are not directly clinically relevant, as they were all under 0.5 mm. Furthermore, a small difference in the patient’s face will occur in between image acquisitions. This is known to be around 0.25 mm average for sequential images, and 0.36 mm for images with a time difference of six weeks.19 However, these differences will be present for every imaging system. This indicates that 3D images of those three systems are comparable with each other in terms of accuracy, and that these systems can be used simultaneously in, for example, multicenter studies. CONCLUSION The 3dMDface system and the Vectra XT have similar reproducibility, while the Artec Eva has significantly lower reproducibility. The Vectra XT and Artec Eva showed a small mean difference with the 3D camera from 3dMD. However, the differences found were not clinically relevant, indicating that 3D images of all three systems are comparable and can be used simultaneously in multicenter studies. The main advantage of the Artec Eva is that it is a mobile scanner, while the 3dMDface system and Vectra XT are rigid systems that cannot be transported easily. Also, the costs of the Artec Eva are significantly lower than the costs of the 3dMDface system and Vectra XT. However, the Artec Eva requires more acquisition time and is therefore more prone to movement artifacts. Furthermore, the data processing for the Artec Eva is performed manually. In contrast, the 3dMDface system and Vectra XT have fast acquisition times and have 3D images almost instantly available for analysis and surgical planning. Supplementary Material This article contains supplementary material located online at www.aestheticsurgeryjournal.com. Disclosures The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. Funding The authors received no financial support for the research, authorship, and publication of this article. The authors declared no potential conflicts of interest with respect to the research, authorship, and publication of this article. REFERENCES 1. Meulstee J , Liebregts J , Xi T et al. A new 3D approach to evaluate facial profile changes following BSSO . J Craniomaxillofac Surg . 2015 ; 43 ( 10 ): 1994 - 1999 . Google Scholar CrossRef Search ADS PubMed 2. van Loon B , Maal TJ , Plooij JM et al. 3D Stereophotogrammetric assessment of pre- and postoperative volumetric changes in the cleft lip and palate nose . Int J Oral Maxillofac Surg . 2010 ; 39 ( 6 ): 534 - 540 . Google Scholar CrossRef Search ADS PubMed 3. Hopman SMJ , Merks JHM , Suttie M , Hennekam RCM , Hammond P . Face shape differs in phylogenetically related populations . Eur J Hum Genet . 2014 ; 22 ( 11 ): 1268 - 1271 . Google Scholar CrossRef Search ADS PubMed 4. Baan F , Liebregts J , Xi T et al. A New 3D Tool for Assessing the Accuracy of Bimaxillary Surgery: The OrthoGnathicAnalyser . PLoS One . 2016 ; 11 ( 2 ): e0149625 . Google Scholar CrossRef Search ADS PubMed 5. Hameeteman M , Verhulst AC , Maal TJJ , Ulrich DJO . An analysis of pose in 3D stereophotogrammetry of the breast . J Plast Reconstr Aesthetic Surg . 2016 ; 69 ( 12 ): 1609 - 1613 . Google Scholar CrossRef Search ADS 6. Hameeteman M , Verhulst AC , Vreeken RD , Maal TJJ , Ulrich DJO . 3D stereophotogrammetry in upper-extremity lymphedema: An accurate diagnostic method . J Plast Reconstr Aesthetic Surg . 2016 ; 69 ( 2 ): 241 - 247 . Google Scholar CrossRef Search ADS 7. 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 ( 3 ): 487 - 494 . Google Scholar CrossRef Search ADS PubMed 8. Plooij JM , Swennen GRJ , Rangel FA et al. Evaluation of reproducibility and reliability of 3D soft tissue analysis using 3D stereophotogrammetry . Int J Oral Maxillofac Surg . 2009 ; 38 ( 3 ): 267 - 273 . Google Scholar CrossRef Search ADS PubMed 9. Lübbers H-T , Medinger L , Kruse A , Grätz KW , Matthews F . Precision and Accuracy of the 3dMD Photogrammetric System in Craniomaxillofacial Application . J Craniofac Surg . 2010 ; 21 ( 3 ): 763 - 767 . Google Scholar CrossRef Search ADS PubMed 10. 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 ( 3 ): 477 - 483 . Google Scholar CrossRef Search ADS PubMed 11. Metzler P , Sun Y , Zemann W et al. Validity of the 3D VECTRA photogrammetric surface imaging system for cranio-maxillofacial anthropometric measurements . Oral Maxillofac Surg . 2014 ; 18 ( 3 ): 297 - 304 . Google Scholar CrossRef Search ADS PubMed 12. Spanholtz T , Leitsch S , Holzbach T , Volkmer E , Engelhardt T , Giunta R . 3-dimensionale Bilderfassung: Erste Erfahrungen in der Planung und Dokumentation plastisch-chirurgischer Operationen . Handchirurgie Mikrochirurgie Plast Chir . 2012 ; 44 ( 4 ): 234 - 239 . Google Scholar CrossRef Search ADS 13. Modabber A , Peters F , Kniha K et al. Evaluation of the accuracy of a mobile and a stationary system for three-dimensional facial scanning . J Craniomaxillofac Surg . 2016 ; 44 ( 10 ): 1719 - 1724 . Google Scholar CrossRef Search ADS PubMed 14. Maal TJJ , van Loon B , Plooij JM et al. Registration of 3-dimensional facial photographs for clinical use . J Oral Maxillofac Surg . 2010 ; 68 ( 10 ): 2391 - 2401 . Google Scholar CrossRef Search ADS PubMed 15. 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 ( 3 ): 247 - 253 . Google Scholar CrossRef Search ADS PubMed 16. Metzger TE , Kula KS , Eckert GJ , Ghoneima AA . Orthodontic soft-tissue parameters: a comparison of cone-beam computed tomography and the 3dMD imaging system . Am J Orthod Dentofacial Orthop . 2013 ; 144 ( 5 ): 672 - 681 . Google Scholar CrossRef Search ADS PubMed 17. De Menezes M , Rosati R , Ferrario VF , Sforza C . Accuracy and reproducibility of a 3-dimensional stereophotogrammetric imaging system . J Oral Maxillofac Surg . 2010 ; 68 ( 9 ): 2129 - 2135 . Google Scholar CrossRef Search ADS PubMed 18. Maal TJJ , Plooij JM , Rangel FA , Mollemans W , Schutyser FAC , Bergé SJ . The accuracy of matching three-dimensional photographs with skin surfaces derived from cone-beam computed tomography . Int J Oral Maxillofac Surg . 2008 ; 37 ( 7 ): 641 - 646 . Google Scholar CrossRef Search ADS PubMed 19. Maal TJJ , Verhamme LM , van Loon B et al. . Variation of the face in rest using 3D stereophotogrammetry . Int J Oral Maxillofac Surg . 2011 ; 40 : 1252 - 1257 . © 2018 The American Society for Aesthetic Plastic Surgery, Inc. Reprints and permission: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Aesthetic Surgery JournalOxford University Press

Published: Jan 18, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off