Introduction

Facial beauty is a long-standing mystery in human history. Numerous scholars have investigated what facial beauty is, and whether it can be defined objectively. While artists have expressed their own concepts of facial beauty in their work, mathematicians and scientists have tried to analyze beauty and determine its rules and definitions [1, 2]. For example, the golden ratio, also known as the divine proportion, is considered by many to hold the key to the mystery of esthetics, attraction, and human beauty [3]. The golden ratio is denoted by the symbol Φ (phi) and is an irrational number near 1.618033988 [4].

Although all human beings have similarly shaped eyes, noses, and lips, each individual looks very different, even twins. Which factors create these differences between people? Although subjective preferences in facial beauty exist, commonalities are also known to exist, regardless of race, age, and sex. According to recent studies, facial beauty can be defined using three components: facial balance, youthfulness, and symmetry. Among these three factors, recent studies have found that differences between individuals are due to differences in facial balance, such as proportional differences in various facial landmarks.

Many investigators have used photography for this purpose. However, two-dimensional (2D) analyses of the face have a potential inevitable measurement bias. Three-dimensional (3D) analyses enable us to overcome this obstacle.

In this study, a comparison of facial proportions was performed between beauty pageant contestants and ordinary young women with Korean ethnicity using 3D photogrammetry. The results of this study may become a fundamental basis for the study of the concept of facial beauty, which would contribute to the clinical application of this concept for plastic surgeons.

Materials and Methods

All participants underwent 3D photography purely for research reasons and were evaluated with 3D photogrammetry. Each participant was photographed with 3D photography between October and November 2016 in Seoul, Korea. In all, 43 beauty pageant contestants (group I) and 48 ordinary young women (group II) were included. Group I was composed of prize-winning contestants in the Miss Korea contest, and group II consisted of patients who visited our outpatient clinic for non-facial treatment. The mean age for group I was 22.3 ± 3 years and the mean age for group II was 25 ± 5 years. Participants with histories of congenital, traumatic, or post-operative facial deformity, those with craniofacial syndromes, and those with severe facial asymmetry were excluded. Medical histories of prior plastic surgery or esthetic procedures were not considered. Informed consent was obtained from all participants, and the study design was reviewed and approved by our institutional review board.

Three-Dimensional Photogrammetry

Using a Morpheus 3D® light-emitting diode-based white structured-light scanner (Morpheus Co., Ltd., Seongnam, Gyunggi-do, Korea), 3D photographs of the participants were taken while they were sitting in a neutral head position, with about a 600 mm distance from the face to the lens of the 3D scanner, without makeup. Each scan was performed with the same brightness, three times in the frontal, left, and right oblique views (Fig. 1). After each participant’s 3D facial image was reconstructed by merging these images, the entire scanned image was automatically reoriented. The horizontal plane was set to contain the right and left pupils and the nasion, the sagittal plane was set as a perpendicular plane containing the nasion and subnasale, and the coronal plane was set to be perpendicular to these other planes. Then, the various facial landmarks of each of the participants were defined. Analyses of facial proportions included absolute lengths, angles, proportions of facial volumes, and vertical and horizontal lengths between each soft tissue facial landmark. More specifically, from each 3D photogrammetric image, after 27 soft tissue landmarks were marked (Table 1, Fig. 2), 13 defined absolute lengths, 5 angles, 3 volumetric proportions, and 12 length proportions from each scanned image were measured and analyzed using a commercial computer software tool (Morpheus Co., Ltd., Seongnam City, Gyunggi-do, Korea) (Table 2, Fig. 3).

Fig. 1
figure 1

The three-dimensional scan was performed three times in the frontal, left, and right oblique views

Full size image
Table 1 Landmarks for facial analysis
Full size table
Fig. 2
figure 2

Landmarks used in this study. See Table 1 for detailed definitions of each landmark. A representative participant of the beauty pageant contestants group (above) and a representative participant of the ordinary young women group (below)

Full size image
Table 2 Definitions of the measures used in the three-dimensional scanned images
Full size table
Fig. 3
figure 3

The Morpheus software tool (Morpheus Inc., Seoul, South Korea) was used for the three-dimensional photogrammetric analyses: volumetric analyses (a), vertical and horizontal length analyses (b), (c), (d), (e), and angular analyses (f) of the face. See Table 2 for detailed definitions of each proportion

Full size image

Statistical Analyses

All statistical analyses were performed using SPSS 24.0 (IBM® SPSS® Statistics, Chicago, IL, USA). Data from each group were analyzed to determine the mean P value as well as the standard deviation of each measurement. The between-group mean differences were calculated using t test, with the significance level set to a P value of 0.05.

Results

Anthropometric analyses of facial parameters are shown in Table 3. Average values ± standard deviations are discussed in the following section. There were statistically significant differences between the beauty pageant contestants and ordinary young women in 18 of the 33 measures taken.

Table 3 Comparison of the variables between group I (beauty pageant contestants) and group II (ordinary young women)
Full size table

Absolute Lengths and Angles

Middle face height was significantly longer (P = 0.043) in the beauty pageant contestants (72.36 ± 3.65 mm) than the ordinary young women (70.8 ± 3.55 mm). By contrast, lower facial height was significantly shorter (P = 0.002) in the beauty pageant contestants (60.26 ± 2.53 mm) than the ordinary young women (62.61 ± 4.26 mm). In terms of the nose, the vertical height of the nose was significantly higher (P = 0.000) in the beauty pageant contestants (54.56 ± 2.84 mm) than the ordinary young women (51.87 ± 2.74 mm), but the horizontal width of the nose was significantly shorter (P = 0.001) in the beauty pageant contestants (38.05 ± 1.82 mm) than the ordinary young women (39.59 ± 2.29 mm). In terms of the eye, the vertical height of the eye was significantly higher (P = 0.000) in the beauty pageant contestants (11.55 ± 1.03 mm) than the ordinary young women (9.14 ± 1.27 mm). The horizontal width of the eye was also larger (P = 0.000) in the beauty pageant contestants (28.01 ± 1.22 mm) than the ordinary young women (26.18 ± 1.6 mm). By contrast, the intercanthal width was significantly shorter (P = 0.001) in the beauty pageant contestants (34.36 ± 1.86 mm) than the ordinary young women (36.28 ± 3.16 mm) (Fig. 4). There were no significant differences between the groups in the vertical height of total face, upper face, chin, lip, or in the horizontal width of the total face and lip.

Fig. 4
figure 4

Comparison of the mean absolute length values with statistically significant differences. The beauty pageant contestants group had greater middle face length, nasal height, eye height, and eye width; they also had a smaller lower face height, intercanthal width, and nasal width than the ordinary young women group

Full size image

In terms of facial angles, only the nasolabial angle was larger in the beauty pageant contestants (118.4° ± 6.27°) than ordinary young women (108.32° ± 8.79°). This difference was statistically significant (P = 0.000). None of the other angle variables were statistically significantly different (Fig. 5).

Fig. 5
figure 5

Comparison of the mean facial angle values with statistically significant differences. Only the nasolabial angle was significantly different (larger) the beauty pageant contestants group compared with the ordinary young women group (asterisk)

Full size image

Volumetric Proportions

The ratio between upper facial volume and total facial volume showed a significant difference between the two groups (beauty pageant contestants, 14.91 ± 2.01%; ordinary young women, 13.76 ± 2.39%; P = 0.014). The ratio between middle facial volume and total facial volume also showed a significant difference between the two groups (beauty pageant contestants, 52.96 ± 2.20%; ordinary young women, 51.33 ± 2.28%; P = 0.001). In terms of the ratio between lower facial volume and total facial volume, the beauty pageant contestants had a significantly lower proportion than the ordinary young women (beauty pageant contestants, 32.11 ± 2.42%; ordinary young women, 34.90 ± 3.16%; P = 0.000) (Fig. 6).

Fig. 6
figure 6

Comparison of the mean volumetric proportion values with statistically significant differences. The beauty pageant contestants group had a larger proportional volume of the upper and middle face than the ordinary young women group, and had a smaller proportional volume of the lower face

Full size image

Length Proportions

As described above, the proportions of all vertical heights were calculated as a ratio to the total vertical length of the face, and the proportions of all horizontal widths were calculated as a ratio to the total width of the face. The proportion of lower face was significantly smaller (P = 0.000) in the beauty pageant contestants (32.12 ± 1.27%) than the ordinary young women (33.46 ± 1.64%). There were no differences between the two groups in the proportions of facial index, upper, and middle facial height, and chin height. In terms of the proportions of nose, the proportion of nasal vertical height was significantly greater (P = 0.000) in the beauty pageant contestants (29.08 ± 1.42%) than the ordinary young women (27.76 ± 1.61%). By contrast, the proportion of nasal horizontal width was significantly smaller (P = 0.000) in the beauty pageant contestants (29.33 ± 1.44%) than ordinary young women (31.01 ± 2.11%). In terms of the eye, both vertical height proportion and horizontal width proportion were significantly greater in the beauty pageant contestants (6.15 ± 0.54 and 21.58 ± 0.69%, respectively) than the ordinary young women (4.89 ± 0.7 and 20.48 ± 0.99%) (P = 0.000 and 0.000, respectively). The proportion of intercanthal width was significantly smaller (P = 0.000) in the beauty pageant contestants (26.47 ± 1.15%) than the ordinary young women (28.36 ± 1.91%). Although the proportion of horizontal mouth width showed a significant difference between the two groups (beauty pageant contestants, 33.88 ± 2.11%; ordinary young women, 34.78 ± 2.16%), there was no difference between the two groups in the proportion of vertical mouth height (beauty pageant contestants, 9.24 ± 1.05%; ordinary young women, 8.83 ± 1.25%) (Fig. 7).

Fig. 7
figure 7

Comparison of the mean values of proportions of length with statistically significant differences. The beauty pageant contestants group had a greater nasal height, eye height, and eye width than the ordinary young women group, and a smaller lower face height, intercanthal width, nasal width, and mouth width

Full size image

Discussion

The concept of facial beauty proposed by Leonardo da Vinci is the standard concept in modern esthetic plastic surgery [5]. It hypothesizes that the golden ratio is ideal, even in facial beauty. This concept has played a key role in the innovation of surgical techniques, including in blepharoplasty, rhinoplasty, and so on. Thus, the golden ratio classically depends on the concept of facial proportion. Why do people prefer the facial beauty of certain individuals? Which factors contribute to these preferences? To help answer these questions, we obtained 3D images of prize-winning contestants in the Miss Korea contest to represent attractive women. Because the prize-winning standards of beauty pageants not only focus on the face, but also on body shape, with body mass index or even overall body ratios taken into account, it may be ambitious to conclude that beauty pageant standards reflect the absolute facial beauty standard. Although it is disputed whether Miss Korea contestants represent ideal facial beauty, considering the fact that the judges include not only experts such as plastic surgeons, orthodontists, and photographers, but also laypeople such as journalists and celebrities, we believe that Miss Korea contestants provide an appropriate approximation of ideal beauty that could be considered as a facial beauty standard.

Many ordinary young women in northeastern Asia would prefer to have a smaller and slenderer facial contour. Why? This preference relates to the proportions between the face and the body. Asians are somewhat smaller in stature than Western people, so their faces are somewhat larger in proportion to their bodies [6]. In addition, contemporary Asian facial contouring surgery often deals with mandible reduction, which is a procedure to change the proportions of the lower third of the face [7,8,9]. In the present study, participants in the beauty pageant contestants group had smaller absolute length, volumetric proportion, and length proportion in their lower faces than the ordinary young women.

There are generally three methods used for facial analysis: cephalometry, direct anthropometry, and 2D photogrammetry (indirect anthropometry). Each of these methods has deadly flaws and pitfalls [10]. Cephalometry is generally a valuable technique when used for surgical planning and evaluation of pre- and post-operative outcomes. However, its validity, sensitivity, and significance are matters of ongoing dispute. Cephalometric normative data, for example, are not representative of facial attractiveness [11]. Anthropometry is also widely used in clinical settings. However, it is a very laborious procedure and is subject to some error. Currently, 2D photogrammetry is the most widely used method. However, it offers only a limited number of reliable measurements, due to distortion and the 2D nature of the photographs [12,13,14]. This is particularly true for frontal views. Thus, we used 3D photogrammetry, because it is a relatively error-free and objective method [15, 16]. It does not suffer from the limitations listed above. Indeed, our results indicate that this method was reliable and stable for evaluating facial beauty.

There have been various subjective and objective studies on standards of facial attractiveness, and numerous studies on the differences among races, between the sexes, and due to changing times [17, 18]. Rhee et al. [10] tried to develop a new digital photogrammetric method of facial analysis known as balanced angular and proportional analysis. However, that study was based on 2D photogrammetry and focused on sex differences. There have been very few studies that have compared facial proportions between ordinary young women and beauty pageant contestants of Asian ethnicity, based on 3D photogrammetry.

Although the purpose of reconstructive surgery can be achieved by creating an average face, the aim of esthetic surgery is to create an attractive face. Most studies have proposed the average face as a standard, or compared proportional differences between pre- and post-operative measures, assuming that all post-operative results have attractive proportions. However, we have tried to present the true standard of an attractive face by comparing prize winners in the Miss Korea contest, who are generally recognized as beautiful young women, to ordinary young women.

Of course, the proportions of facial volumes and lengths are more important than absolute values [19]. However, in practice, it is undeniable that the proportion of the face and the absolute length of the facial structure are invaluable when evaluating a patient’s facial profile, communicating with a patient, or performing surgery. For this reason, we have included information on absolute facial lengths as well as facial proportions. Our results document differences between beauty pageant contestants and ordinary young women. Beauty pageant contestants have a greater length of the middle face, nasal height, and eye height and width, and a smaller length of lower face, intercanthal width, and nasal width. They also have a smaller proportion of lower facial volume and height, intercanthal width, nasal width, and mouth width. Further, they have a larger proportion of eye width and height, nasal height, and nasolabial angle. These results remind us of the importance of facial proportions (Fig. 8).

Fig. 8
figure 8

a Beauty pageant contestants had a greater middle face length, nasal height, and eye height and width, and a larger nasolabial angle (above), whereas ordinary young women had a greater lower face length, intercanthal width, and nasal width (below). b Beauty pageant contestants had a larger proportion of eye width, height, nasal height, and upper and middle facial volume (above). By contrast, ordinary young women had a larger lower facial volume and height, intercanthal width, nasal width, and mouth width (below)

Full size image

Finally, it is possible that the average values of the profiles of the Miss Korea contestants listed in Table 3 could be used to give approximations for attractive facial profile for young women. This would be useful information when performing esthetic and reconstructive surgeries in clinical settings (Fig. 9).

Fig. 9
figure 9

Ideal facial profile created using the 3D photogrammetric analyses of images of prize-winning Miss Korea contestants

Full size image

Conclusions

This 3D photogrammetric study has revealed that there are differences between the faces of ordinary young women and beauty pageant contestants, and has identified the factors that contribute to facial beauty. This is the first investigation of facial beauty based on 3D photogrammetric analyses of ordinary young women and beauty pageant contestants in East Asia.