Influence of facial width on the perception of lip protrusion and its differences between profile and 3-dimensional video clip

Introduction

This study evaluated the influence of facial width on the perception of lip protrusion and investigated the concordance between 2-dimensional (2D) profile images and 3-dimensional (3D) video clips in assessing lip protrusion.

Methods

An Asian female standard head model was created using 3D modeling software. Eight head models were constructed by modifying the standard head model in terms of facial width (broad, neutral, and slim) and lip protrusion (retrusive, straight, and protrusive). Overall, 97 Asian raters rated the lip protrusion from the 2D profiles and 3D rotation video clips of the 9 models.

Results

No significant differences were found in the perception of lip protrusion in terms of sex, age, or occupation. Compared with the 2D profiles, the 3D video clips were rated as more protrusive in 8 of the 9 head models, with the retrusive broad, retrusive neutral, straight broad, and straight slim faces showing statistical significance ( P <0.01). The rating is significantly higher in slim faces than in broad faces across the 3 groups of 2D profiles ( P <0.01). For 3D video clips, the rating was higher in slim faces than in broad faces in all 3 groups, whereas differences were significant in the straight and protrusive groups only ( P <0.01).

Conclusions

In this study, 3D video clips were more sensitive to the perception of lip protrusion than were 2D profiles to some extent. The lips were rated relatively more protrusive in a slim face than in a broad face. Therefore, the relationship between facial width and lip protrusion should be considered in orthodontic treatment goals and treatment plans.

Highlights

•

3D videos were more sensitive in lip protrusion perception than 2D profiles to some extent.
•

The lips were rated relatively more protrusive in a slim face than in a broad face.
•

Facial width may influence the perception of lip protrusion.

Facial morphology and proportion are quantifiable and objective ^, ; however, the perception and evaluation of facial esthetics are complex and subjective, which are influenced by combined factors, including anatomic characteristics, ethnicity and culture, sex and age, educational background, ^, and psychosocial factors. Most studies evaluating lip protrusion were based on 2-dimensional (2D) images. As far as orthodontists are concerned, 2D images such as portraits, profiles, and lateral cephalograms are adopted as the most classic and efficient carriers for assessing facial attractiveness. Regarding the harmony and balance between the lips, nose, chin, and teeth, a certain degree of consensus and relatively uniform standards exist. ^,

The human face is a complex 3-dimensional (3D) structure that people may perceive from different angles in addition to the frontal and lateral views, which means that 2D photographs probably lose cues or information compared with 3D images. With the development of 3D imaging techniques, 3D images that combine multiple view angles are employed for assessing the correlation between facial attractiveness and dentoskeletal relation, ^, cheek volume, facial asymmetries, smile dimensions, etc. Instead of dynamic 3D images, static 3D images with some specific view angles were used in most of these studies. ^, Although several studies have compared perceptual differences between 2D and 3D images, ^, the relevant literature is limited, and the results of the comparison are still being debated.

Sagittal facial esthetics has always been the priority in orthodontics, followed by vertical issues. As the lips are considered the focal point of lower face esthetic concern, numerous evaluations on the harmonious relationship between lip protrusion and anteroposterior facial types have been reported. ^, ^, In addition, a few studies aimed to discover the most desirable lip position for different vertical facial disproportions. ^, The favored lip position differed with the diverse anteroposterior and/or vertical facial types. A study reported that the mean preferred chin projection angle differed significantly according to the front facial index, which is the ratio of the facial height and facial width. However, the literature on the perception of lip attractiveness, according to transversal facial types (ie, facial width), is still lacking.

This study aimed to (1) investigate whether there are any differences in the perception of lip protrusion between 2D and 3D images and (2) identify the effect of facial width on the perception of lip protrusion. The null hypothesis was that the preference for lip protrusion would not be influenced by facial width and that no differences would be found in the perception of lip protrusion between 2D and 3D images.

Material and methods

This study was conducted in accordance with the Declaration of Helsinki and was approved by the institutional review board of Kyungpook National University Dental Hospital (no. KNUDH-2023-05-01-00).

An Asian female volunteer with ideal occlusion, balanced facial soft tissue, and normal Asian cephalometric values was selected. Using the 3D modeling software (version 3.0.1; Blender Foundation, Amsterdam, Netherlands), an original Asian female head model was created according to the volunteer’s facial and head features ( Fig 1 ). The facial index (facial height/facial width) = 0.875, which fell within the category of a mesoprosopic facial type. The tips of the upper and lower lips fell on the Ricketts E line to ensure a straight profile. Then, the head model was modified in terms of facial width and/or lip protrusion. Specifically, bizygomatic width and bigonial width were increased or decreased by 5% (broad, 5% and slim, −5%). Simultaneously, there were appropriate adjustments in the fullness of the inferior area of zygomatic arches and cheeks in 3 dimensions, rendering it excessively natural and conforming to the facial width ( Fig 2 , A ). The upper and lower lips were morphed backward to 3 mm behind the E line for a retrusive profile or forward to 3 mm beyond the line for a protrusive profile ( Fig 2 , B ). The modification of facial width and lip protrusion did not interfere with each other. These processes were performed by an experienced professional. As a result, another 8 head models were created. All 9 face models were defined as the protrusive broad, protrusive neutral, protrusive slim, straight broad, straight neutral (original face), straight slim, retrusive broad, retrusive neutral, and retrusive slim faces ( Fig 3 ).

These 9 models were classified into 3 groups: protrusive, straight, and retrusive groups. Each group included 3 models of the same lip protrusion and different facial widths. Lateral views (90°) of the 9 models were captured as 2D profiles. For 3D views, a clip was recorded for each head model, in which the head started from the oblique view and rotated around the y-axis once to the original position. Each clip lasted 12 seconds (See Video , available at www.ajodo.org ).

In total, the study analyzed 97 Asian raters, including laypersons, orthodontists, and dentists. Information about age, sex, and occupation or major was collected. All individual raters were well-informed about lip protrusion but had no information concerning the modification of the images. Participants with visual impairment, aged <18 years, or having a psychiatric condition were excluded. All raters were asked to rate each profile and video clip in 2 sessions of evaluation for lip protrusion on the basis of intuition. In the first session, the 2D profiles of the models were arranged randomly and placed side by side in each group. Figure 4 shows an example of the arrangement in the retrusive group. In the second session, 3D rotation clips were arranged randomly and placed side by side in each group.

A 10-point Likert scale (1, very retrusive and 10, very protrusive) was used. Giving the same rating was allowed if the lips were supposed to be the same protrusion. There was no time limit for rating. After 4 weeks, 10 randomly selected raters (every 10th rater starting from rater 5 to 95) were asked to do the evaluation again to assess the intrarater reliability.

Statistical analysis

SPSS software (version 26.0; IBM, Armonk, NY) was used for statistical analysis. Data were checked for normal distribution by the Kolmogorov-Smirnov test. As a result, the survey response data were not normally distributed. Spearman’s rank correlation coefficient was used for the intrarater reliability test. Mann-Whitney U tests were used to evaluate the difference in sexes, whereas the Kruskal-Wallis analyses followed by multiple comparisons post-hoc tests were performed for comparison within occupations. The Wilcoxon signed rank test was used to compare the differences in perception between 2D profile images and 3D video clips for each modification. The Friedman test with Bonferroni correction was used to detect the differences in perception between modifications of the same lip protrusion and different facial widths.

Results

A total of 97 Asian raters participated in the survey. The proportion of males (n = 32, 33.0%) and females (n = 65, 67.0%) was approximately 1:2. Most participants were laypersons (n = 39, 40.2%), followed by orthodontists (n = 34, 35.1%) and dentists (n = 24, 24.7%). Most were aged 18-40 years (n = 93, 95.9%) ( Table I ). Of the 10 re-raters, 5 were males and 5 were females, with 5 orthodontists, 1 dentist, and 4 laypersons. The intrarater reliability was moderate to very good ( r _s = 0.533-0.936), with the highest r _s observed for a female orthodontist and the lowest for a male dentist.

Table I

Characteristics of 97 raters in the survey

Characteristics	n (%)
Sex
Male	32 (33.0)
Female	65 (67.0)
Age, y
18-30	52 (53.6)
31-40	41 (42.3)
41-50	1 (1.0)
51-60	1 (1.0)
>60	2 (2.1)
Major
Orthodontist	34 (35.1)
Dentist	24 (24.7)
Layperson	39 (40.2)

Table II shows no significant differences between sexes except for the rating of the straight, broad face. No significant differences were observed among orthodontists, dentists, and laypersons, except for the rating of the protrusive neutral face.

Table II

Comparison between sexes and majors

Variables	Sex		Major
Variables	Male	Female	Orthodontist	Dentist	Layperson
2D profiles
Retrusive broad	3.50 ± 1.61 (4.0, 2.0-5.0)	3.77 ± 1.44 (4.0, 2.5-5.0)	3.74 ± 1.26 (4.0, 3.0-5.0)	3.80 ± 1.45 (4.0, 3.0-5.0)	3.59 ± 1.73 (4.0, 2.0-5.0)
Retrusive neutral	3.94 ± 1.85 (4.0, 2.0-5.0)	4.12 ± 1.57 (5.0, 3.0-5.0)	3.79 ± 1.43 (4.0, 3.0-5.0)	4.17 ± 1.31 (5.0, 3.0-5.0)	4.23 ± 2.01 (5.0, 2.0-5.0)
Retrusive slim	4.80 ± 2.08 (5.0, 3.0-6.0)	4.68 ± 2.05 (5.0, 3.0-6.0)	4.26 ± 1.60 (4.0, 3.0-5.3)	5.46 ± 2.21 (5.5, 4.3-7.0)	4.62 ± 2.22 (5.0, 3.0-6.0)
Straight broad	4.56 ± 1.74 (5.0, 4.0-5.8)	4.46 ± 1.65 (4.0, 3.0-5.0)	4.47 ± 1.24 (5.0, 4.0-6.0)	4.63 ± 1.79 (4.0, 4.0-5.0)	4.44 ± 1.94 (5.0, 3.0-6.0)
Straight neutral	5.13 ± 1.07 (5.0, 5.0-6.0)	4.72 ± 1.39 (5.0, 4.0-6.0)	4.88 ± 1.17 (5.0, 4.0-6.0)	4.92 ± 1.18 (5.0, 4.3-5.8)	4.79 ± 1.49 (5.0, 4.0-6.0)
Straight slim	5.59 ± 1.86 (6.0, 5.0-6.8)	5.02 ± 1.63 (5.0, 4.0-6.0)	5.47 ± 1.48 (6.0, 5.0-6.0)	5.29 ± 1.92 (5.0, 5.0-6.0)	4.92 ± 1.80 (5.0, 4.0-6.0)
Protrusive broad	6.44 ± 1.98 (7.0, 5.0-8.0)	6.46 ± 1.81 (7.0, 5.0-8.0)	6.65 ± 1.72 (7.0, 6.0-8.0)	6.58 ± 1.86 (7.0, 5.3-8.0)	6.21 ± 1.99 (6.0, 5.0-8.0)
Protrusive neutral	7.09 ± 1.94 (7.0, 6.3-8.8)	6.83 ± 1.93 (7.0, 5.5-8.0)	7.15 ± 1.76 (7.0, 5.0-8.0) ^ab	7.46 ± 1.96 (7.5, 7.0-9.0) ^a	6.38 ± 1.97 (7.0, 5.0-8.0) ^b
Protrusive slim	7.38 ± 1.56 (8.0, 7.0-8.0)	7.17 ± 1.94 (7.0, 6.0-9.0)	7.41 ± 1.73 (8.0, 6.8-9.0)	7.50 ± 1.67 (7.5, 7.0-9.0)	6.92 ± 1.98 (7.0, 5.0-8.0)
3D video clips
Retrusive broad	4.63 ± 2.09 (5.0, 3.0-5.0)	4.58 ± 1.97 (5.0, 3.0-5.0)	4.06 ± 1.74 (4.0, 3.0-5.0)	5.08 ± 2.02 (5.0, 4.0-6.0)	4.77 ± 2.13 (5.0, 4.0-6.0)
Retrusive neutral	4.44 ± 1.52 (5.0, 4.0-5.8)	4.58 ± 1.64 (5.0, 4.0-5.0)	4.32 ± 1.49 (4.0, 3.0-5.0)	4.83 ± 1.69 (5.0, 4.0-5.8)	4.54 ± 1.64 (5.0, 4.0-5.0)
Retrusive slim	4.69 ± 1.69 (5.0, 4.0-5.0)	4.80 ±1.97 (5.0, 4.0-6.0)	4.71 ± 1.47 (5.0, 4.0-6.0)	4.21 ± 1.53 (5.0, 3.3-5.0)	5.08 ± 2.30 (5.0, 4.0-6.0)
Straight broad	5.63 ± 1.21 (6.0, 5.0-6.0)	4.95 ±1.30 (5.0, 5.0-6.0) ^∗	5.18 ± 1.00 (5.0, 5.0-6.0)	5.29 ± 1.33 (5.0, 5.0-6.0)	5.10 ± 1.54 (5.0, 4.0-6.0)
Straight neutral	5.22 ± 1.84 (5.0, 4.0-6.0)	5.00 ± 1.82 (5.0, 4.0-6.0)	5.35 ± 1.07 (5.0, 4.8-6.0)	4.50 ± 1.64 (5.0, 4.0-6.0)	5.18 ± 2.34 (5.0, 4.0-7.0)
Straight slim	6.13 ± 1.72 (6.0, 5.0-7.0)	5.82 ± 1.79 (6.0, 5.0-7.0)	6.00 ± 1.41 (6.0, 5.0-7.0)	6.08 ± 1.69 (6.0, 5.0-6.8)	5.74 ± 2.09 (6.0, 5.0-7.0)
Protrusive broad	6.44 ± 1.78 (7.0, 5.3-8.0)	6.20 ± 1.99 (6.0, 5.0-7.0)	6.76 ± 1.58 (7.0, 6.0-8.0)	6.38 ± 2.20 (7.0, 5.3-8.0)	5.79 ± 1.94 (6.0, 4.0-7.0)
Protrusive neutral	7.59 ± 1.81 (8.0, 6.0-9.0)	7.15 ± 1.70 (7.0, 6.0-8.0)	7.56 ± 1.64 (8.0, 7.0-8.3)	7.79 ± 1.69 (8.0, 6.0-9.0)	6.77 ± 1.8 (7.0, 5.0-8.0)
Protrusive slim	7.41 ± 1.68 (8.0, 7.0-9.0)	7.40 ± 1.72 (8.0, 6.0-9.0)	7.74 ± 1.58 (8.0, 7.0-9.0)	7.58 ± 1.44 (8.0, 7.0-8.8)	7.00 ± 1.89 (7.0, 6.0-8.0)