Frontal And Lateral Facial Attractiveness:Is There A Correlation?

David Avila, Zlatko Devcic BS, Koohyar Karimi BS, Brian J.F. Wong, MD, PhDDepartment of Otolaryngology – Head and Neck Surgery, University of California Irvine

OBJECTIVES:This study will (1) determine the relationship between anterior (AP) and lateral (LAT) facial attractiveness and (2) identify anatomic features that may lead to discordance between frontal and lateral facial attractiveness.

INTRODUCTION:The parameters that define frontal and lateral facial attractiveness have been well-defined and delineated in countless text. Despite immense work focused on rigorously quantify metrics that define facial attractriveness, to the best of our knowledge, no study has examined whether a correlation exists between the anterior (AP) and lateral (LAT) facial attractiveness. This study will look at the correlation between AP and LAT image attractiveness scores to help characterize facial attractiveness from a new perspective.

METHODS:• 240 frontal and lateral synthetic images were individually posted (Figure 1 A) on an internet-based rating website (Figure 1B).

• Internet-based focus groups were used to rate images individually on an attractiveness scale from 1 to10. Scores stabilize after at least 600 ratings2.

RESULTS:

CONCLUSION:• We found a stronger correlation between frontal and lateral facial images in the average to more attractive faces, suggesting that facial attractiveness may be consistently easier to identify than unattractiveness.

• Furthermore, specific facial landmarks that lead to greater discordance between frontal and lateral facial attractiveness scores were identified, suggesting that the correction of these landmarks may increase facial harmony and attractiveness.

References:1)Baudouin, J. Y., Tiberghien, G. 2004. Symmetry, averageness, and feature size in the facial attractiveness of women. Acta Physiol. 117:313-332.2) Devcic, Zlatko, Koohyar Karimi, and Brian J.F. Wong. "A Web-based Method for Rating Facial Attractiveness." The Laryngoscope 120.5 (2010): 902-06. Print.3) Langlois, J. H., Roggman, L. A. 1990. Attractive faces are only average. Psychological Science. 1:115-121. 4) Rhodes, G., Yoshikawa, S., Clark, A., Lee, K., McKay, R., Akamatsu, S. 2001. Attractiveness of facial averageness and symmetry in non-western cultures: in search of biologically based standards of beauty. Perception. 30:611-625

7.2 Attractive Frontal and Lateral Attractive frontal and unattractive lateral

Unattractive frontal and attractive lateral Unattractive frontal and unattractive lateral

Figure 1. (A) Representative images from pool of synthetic AP and LAT facial images. (B) individual accounts (hotornot.com) created for each image to allow evaluation and scoring. (C) AP and LAT as it appears on the website. (D) Image as it appears during the rating process with attractiveness score and histogram.

(B)

(A)

Figure 5. (A) Distribution of AP and LAT scores. (B) Percent difference between AP and LAT images in comparison to AP attractiveness scores.

Red: LAT, Blue: AP

Figure 3. Strongly correlated AP and LAT images with increasing attractiveness scores from left to right (Range: 3.3-9.5).

Increasing Attractiveness Score3.4 5.2 6.3 7.6 8.5

3.3 5.1 6.1 7.6 8.4

9.5 9.4 7.2

9.5 7.4 4.5 4.7

7.9 6.5

Frontal Lateral

Size of the lips Blunt cervicomental angle

Asymmetrical features (brows) Long lower third

A square shape, masculine affect Protruding chin

Bulbous nasal tip Broad nasal dorsum

Nasojugal fold Dorsal nasal hump

Figure 6. Several AP/LAT pairs analyzed for discordance of facial features.

4.6 5.2 5.14.1

6.5 5.7

• Additional qualitative analysis was performed, by expert evaluators, on paired outliers where the frontal and lateral attractiveness scores deviated by a percent difference of 20 or greater.

Figure 4. The images are examples of AP and LAT images with either a weak or strong correlation. (A) and (D) demonstrate a strong correlation, while (B) and (C) demonstrate a weak correlation.

(A) (B)

(C) (D)

(C)

(D)

Set up individual account for each AP and LAT to allow focus group evaluation

• In Figure 2 the attractiveness scores for paired anterior (AP) and lateral (LAT) posterior portraits were plotted to examine whether a strong correlation exist (R=.75).

Figure 2. Correlation between AP and LAT facial image attractiveness scores.

•The distribution of AP and LAT scores according to facial attractiveness scores range: 1. (3-3.9), 2. (4-4.9), 3. (5-5.9),4. (6-6.9), 5. (7-7.9), 6. (8-8.9), and 7. (9-10), (Figure 5 A).

• Pairs with a percent difference greater than 20% were analyzed for abnormal facial features, indicated in red circles (Figure 5 B).

Perc

ent D

iffer

ence

bet

wee

n Fr

onta

ls

and

Late

rals

• Figure 6 shows several AP/LAT pairs containing anatomic features, that lead to the discordance between frontal and lateral attractiveness scores (Table 1).

• Anterior (AP) facial images shown in Figure 3 are paired with their corresponding lateral (LAT) images to depict examples of strongly correlated pairs from the pool of synthetic portraits.

Table 1. Table illustrates different facial landmarks that are responsible for the discrepancies between AP and LAT attractiveness scores.

• The set of images in Figure 4 depict the possible outcomes for AP and LAT pairs having either a strong or weak correlation.

9.5

(A) (B)