Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Comparative facial attractiveness in two and three dimensions
(USC Thesis Other)
Comparative facial attractiveness in two and three dimensions
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
COMPARATIVE FACIAL ATTRACTIVENESS IN TWO AND THREE
DIMENSIONS
by
Jordan Millar, D.M.D
________________________________________________________________
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
CRANIOFACIAL BIOLOGY
May 2007
Copyright 2007 Jordan Millar, D.M.D
ii
Dedication
To my family for their encouragement and support that has enabled me
to achieve my dreams.
iii
Acknowledgements
I would like to thank Dr. James Mah for his guidance throughout the
course of this project. I am indebted to my mother for her time and support, as
well as to Dustin White for his encouragement and creativity. I would like to
thank Dr. Richard Jacobson and his staff for their support in record compilation,
and Chester Wang and Dolphin Imaging for supplying the software and
software support for the project. I would also like to recognize Drs. Glenn
Sameshima and Reyes Enciso for their participation on my committee, Towhid
Salam for his statistical support, and my family for their endless support.
iv
Table of Contents
Dedication ii
Acknowledgements iii
Table of Contents iv
List of Tables v
List of Figures vi
Abstract vii
Chapter 1: Introduction 1
Chapter 2: Review of the Literature 2
Chapter 3: Materials and Methods 24
Chapter 4: Results 31
Chapter 5: Discussion 42
Chapter 6: Conclusion 49
Bibliography 50
Appendix 1 58
Appendix 2 59
v
List of Tables
Table 1. Summary of facial attractiveness studies 5
Table 2. Cephalometric values classifying subjects into three facial profile
types (I, II, III) based on Steiner98 norms 24
Table 3. Subject demographic data 31
Table 4. Rater demographic data 32
Table 5. Comparison of mean score in overall, and by subject and rater
groups 34
Table 6. Multivariate association between dimensionality of image and
score in overall sample and by subject and rater groups 35
Table 7. Association among rater group and score by dimension in overall
sample and by subject groups 36
Table 8. Association among subject group and score by dimension in
overall sample and by rater groups 36
vi
List of Figures
Figure 1. 2-D images (frontal and profile) 25
Figure 2. 3D surface image video clip 26
Figure 3. A flow chart of the image creation process 26
Figure 4. Rank frequencies of Class I females by dimension 37
Figure 5. Rank frequencies of Class II females by dimension 37
Figure 6. Rank frequencies of Class III females by dimension 38
Figure 7. Rank frequencies by laypersons by subject group in 2-D 39
Figure 8. Rank frequencies of orthodontist by subject group in 2-D 39
Figure 9. Rank frequencies of dental student by subject group in 2-D 40
Figure 10. Rank frequencies by laypersons by subject group in 3-D 40
Figure 11. Rank frequencies of orthodontist by subject group in 3-D 41
Figure 12. Rank frequencies of dental student by subject group in 3-D 41
vii
Abstract
Introduction: The purpose of this study was to compare facial esthetic
preferences in two-dimensions (2-D) and three-dimensions (3-D) and to
determine the preferred facial relationship. Methods: Orthodontists (n=33),
dental students (n=26), and laypersons (n=27) assessed 2-D images (frontal,
profile) and 3-D surface images of 20 subject within three facial profile: Class I
(n=9), Class II (n=7), and Class III (n=4) on a visual analog scale. A subset from
each profile type was ranked in order of preference for 2-D and 3-D images
separately. The data was analyzed using univariate and multivariate statistics.
Results: Three-D images are viewed significantly more attractive than 2-D.
Professionals and laypeople preferred the Class I facial profile in 2-D and 3-D.
However, orthodontists rated significantly higher than laypeople and dental
students. Conclusions: Although the clinical significance of these differences is
unknown, 3-D surface imaging is a tool that will bring orthodontic records into
the next generation.
1
Chapter 1: Introduction
Although the primary objective of orthodontic treatment is an esthetic and
functional occlusion, achieving a well-balanced and proportional face is imperative.
The desire to improve facial esthetics is the most common reason people seek
orthodontic treatment.
5,34,47
Therefore, an understanding of facial attractiveness is
essential to the orthodontist when addressing their patients’ needs for improved
facial esthetics.
Recent technological developments have enabled 3-D images to be captured,
creating more realistic and complete record of the patient. The 2-D photographs
historically viewed for rating facial attractiveness are being replaced by 3-D facial
images; yet, the comparison of 2-D and 3-D images has not been fully explored. The
objective of this study was to test the hypothesis that there is no difference between
the profile preference of 2-D and 3-D images in the perception of facial
attractiveness.
Many studies have concluded that the Class I profile is
preferred.
16,18,19,23,44,60,77,83,89,98
The further the deviation is from the Class I profile,
the less attractive the profile is perceived to be.
60,77
However, the perception of
attractiveness may be dependent on the assessor. Professionals are more likely to
prefer a Class I facial profile
than laypersons.
18,19,60,77
In this study, the preferred
facial profiles, in 2-D and 3-D, will be determined by a panel consisting of
professionals and laypersons. In addition, the study will assess the difference in rater
preference.
2
Chapter 2: Review of the Literature
Facial esthetics is not only an important part of orthodontics, but is of interest
to people everywhere. Beauty is defined as “a state of harmony – a balance of facial
proportions – a balanced relationship among skeletal structures, teeth, and soft
tissue”
24
or as the relative measure of balance and harmony.
81
Esthetics is “the study
of beauty and to a lesser extent, its opposite, the ugly.”
71
An individual’s facial appearance is one of his/her most obvious
characteristics and from first appearances, may influences opinions. The importance
of appearance is undeniable and stretches far beyond first impression. Research has
shown it has significant social implications
95
such as social power and success, and
positive influences in all areas of society.
2,25,35,43,96
Not only does beauty affect our
behaviors, but it influences our relationships and therefore mate selection and
reproduction.
65,95,104
A person’s self-perception of facial attractiveness is also critical as it can
affect one’s self-esteem. Miller
67
demonstrated how people with low attractiveness
ratings were perceived negatively and those with higher ratings were judged
significantly more positively.
Perceptions of facial beauty are multifactorial, with genetic, environmental,
and cultural foundations.
70,71
Although some agree that there is cross-cultural
agreement about facial attractiveness,
50,71
others have come to the conclusion that
judgments can be influenced by cultural background.
44,53,56,80,93
3
Methods of facial esthetics evaluation
Improving facial esthetics is the most common reason for seeking orthodontic
treatment.
3,5,10,20,34,47,93,95,105
This is a result of the orthodontist’s ability to alter
dentofacial form and facial balance, and the social implications of esthetics.
Therefore, evaluating faces is an integral part of orthodontic diagnosis and treatment
planning, as outlined by our main goals of treatment, (1) ideal function, (2) ideal
esthetics, (3) maintenance of these ideals.
Esthetics has always been an important part of orthodontics; however, the
path to achieving this goal has changed. Up to the time of Angle, orthodontic
correction was for esthetic reasons only.
40
During Angle’s reign, he changed his
ways and disregarded his early concerns for esthetics and focused only on the teeth.
He felt that this would yield an ideal esthetic result.
4
Ever since Angle’s theory was
proposed, there have been orthodontists who have questioned it, as they believed that
a “normal occlusion” did not necessarily result in a beautiful face. A shift back
towards treating to an esthetic ideal resulted.
Because of orthodontists’ ability to alter dentofacial form, they must have an
understanding of facial beauty and of evaluating facial esthetics. Consequently, the
field of orthodontics has attempted to analyze these features. However, the subjective
nature of esthetics has made it difficult to quantify or to define objectively.
22,65,80
Evaluations for diagnosis and treatment planning are made from both the
clinical examination and the records. The goal of the evaluation is to find facial
disproportions, because disproportionate features are less esthetics than proportionate
4
features.
85
The guidelines used by clinicians today are based on those historically
described in art. Vitruvius divided the face vertically into thirds and Ricketts
88
attempted to correlate Plato’s “golden proportion”
109
to ideal facial proportions.
Moss et al.
69
analyzed the faces of professional models and found they did not fit the
golden proportion; thus, questioning the validity of the application of this ratio to
facial analysis. In spite of this, presently both of these ratios are used in evaluating
faces. In regards to facial symmetry, sub-clinical asymmetries are common in most
people, even among esthetically pleasing faces.
31,73,110
In order to create an esthetic standard, researchers have employed many
different techniques to assess facial attractiveness (Table 1). Methods used to assess
facial profile attractiveness include profile line drawings,
48,83,90
silhouettes,
20,22,23,34,39,52,53
and facial photographs.
7,12,14,18,19,32,41,42,44,46,55,56,60,62,66,72,77-
79,84,93,95,98,100,104,107
A good correlation between live subjects and standardized
photographs validates the use of photographic images for assessment.
41
Orthodontists’ prefer to use objective parameters and therefore have
attempted to quantify facial esthetics by using different morphologic measurements.
The categorizations are based on skeletal, dental, and soft-tissue measurements.
Dental measurements used include overjet
44,46
and Angle
classification.
11,18,19,23,44,45,66,77,79
The latter has only four discrete classes and molar
relationship is not necessarily reflected in the face.
12
Skeletal measurements used
include ANB, SNA, SNB, Wits, U1 to NA, L1 to NB, and Sn-
Gn.
20,46,49,55,56,62,72,77,79,89,93,98,104
A wide range of cephalometric values representing a
5
Table 1. Summary of facial attractiveness studies
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
2-D FACIAL PROFILE VIEW
Riedel,
1950
90
Lateral
cephalograph
profile
16 Untreated
children, 13
adults (n=29)
Untreated vs.
Treated
Orthodontists
(n=72)
1-3
("good,"
"fair,"
"poor")
Good profiles have smaller
ANB (<2.5º), Poor profiles
have larger ANB (>2º).
Agreement.
Greater agreement
as to what is not
good than as to
what is good.
Iliffe,
1960
42
Photographs Female
(n=12)
British laypersons
(n=4355)
Agreement
Martin,
1964
61
Photographs Negro female
(n=10)
1. American Negro
male, 2. American
white male, 3.
Social scientists;
n=115
Rank order American whites and
American blacks share a
common female esthetic
standard - the Caucasian
face.
Agreement
Cox and
van der
Linden,
1971
20
Profile
silhouette
77 male, 77
female
(n=154); Age
18-20,
Caucasian
Attractive vs.
Non-attractive
1. Orthodontists, 2.
Laypersons; n=20
Rank order No significant difference in
cephalometric landmarks
(skeletal or soft-tissue).
Agreement
Foster,
1973
34
Profile
silhouette
Female (n=1);
Age 18;
Caucasian
Altered lips (7
profiles)
1. General dentists,
2. Art students, 3.
Orthodontists, 4.
Black laypersons, 5.
Chinese laypersons,
6. White
laypersons; n=180
Rank order
1-8
(youngest
to oldest)
Fuller profiles: youth and
femininity vs. Straighter
profiles: age and
masculinity
Agreement
6
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Prahl-
Andersen
et al,.
1979
83
Profile line
drawings,
intraoral
photographs
Male (n=1) Altered profile
(11 profiles)
1. Orthodontists, 2.
Dentists, 3. Parents;
n=1276
3 point
Likert scale
N/A Agreement
between
orthodontists and
dentists.
Significant
difference between
professionals and
parents.
De Smit
and
Dermaut,
1984
23
Profile
silhouette
Male (n=1),
Class I
Altered 1. AP, 2.
vertical, 3. nose
dorsum (27
profiles)
Dental students
(n=249)
1-9 Likert
scale
Class I most attractive;
Vertically open profiles
least attractive
Agreement
Burcal et
al., 1987
14
Profile
photographs
(black and
white)
2 male, 2
female (n=4);
Prognathic,
Retrognathic
Altered
mandibular and
lip position (5
profiles)
1. Orthodontists, 2.
Oral surgeons, 3.
Laypersons, 4.
Surgical subjects;
n=254
Difference
vs. no
difference
N/A Agreement
between
orthodontists and
surgeons.
Lundstrom
et al.,
1987
56
Annual
profile
photographs
32 male, 32
female (n=
64); Skeletal
age 12 + 18
Sn-Gn
measurements (3
groups)
1. Orthodontic
faculty, 2. Artists, 3.
Laypersons , 4.
Orthodontic senior
residents, 5.
Orthodontic junior
residents; n=20
1-5 Likert
scale
Horizontal mandibular
growers more attractive
than vertical mandibular
growers. Females more
attractive than males.
Agreement.
Laypersons rated
higher and artists
rater lower.
Lew and
Soh,
1992
52
Profile
silhouette
n=1 Altered AP (5
profiles)
Laypersons seeking
dental treatment in
Singapore
(n=1189); Age
15.3y
1-5 Likert
scale
Orthognathic most
attractive
N/A
7
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Czarnecki
et al.,
1993
22
Profile
silhouette
n=1 Altered 1. Lips,
2. nose, 3. chin
(21 profiles)
Dental professionals
(n=545)
1-7 Likert
scale
Male: straighter vs.
Female: slightly convex
profile. Lip position
dependent on chin and
nose.
N/A
Romani et
al., 1993
93
Profile
photographs
(color)
1 male, 1
female (n=2)
Altered AP and
vertical
proportions
1. Orthodontists, 2.
Laypersons; n=44
N/A Agreement in
horizontal changes.
Laypersons more
sensitive to vertical
changes.
Michiels
and Sather,
1994
66
Profile
photographs
(color),
lateral
cephalograph
White
females
(n=130); Age
18-39
Dental: 1. Class
I, 2. Class II, 3.
Class III
Orthodontists, Oral
surgeons, n=6
Horizontal
(1-4 Likert
scale);
Vertical (1-
4 Likert
scale)
Differences between
photographic and
cephalometric profile
evaluation
Agreement
Cochrane
et al.,
1997
19
Profile
photograph
(black and
white)
2 male, 2
female (n=4);
Class I with
average
vertical facial
proportions
Altered AP and
vertical
1. Caucasian
orthodontist, 2.
Caucasian
laypersons; n=80
1-3 Likert
scale
Class I most attractive;
Class II least attractive.
Sex of subject and rater not
significant.
Agreement within
professional
groups; No
agreement between
professionals and
public.
Professionals more
likely to prefer
Class I profile.
8
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Cochrane
et al.,
1999
18
Profile
photographs
(black and
white)
2 male, 2
female (n=4);
Class I with
average
vertical facial
proportions
Altered AP and
vertical
1. Orthodontists, 2.
Maxillofacial
surgeons, 3. Dental
students, 4. Public;
n=158
Rank order Class I most attractive;
Class II least attractive.
Opposite sex rated more
attractive.
Agreement within
professional
groups; No
agreement between
professionals and
public.
Professionals more
likely to prefer
Class I profile.
Hall et al.,
2000
39
Profile
silhouette
30 African
American, 30
White (n=60)
N/A 1. Orthodontist
(white, African
American), 2.
Laypersons (white,
African Amercian);
n=78
VAS More convex profile and
more prominent lips
preferred on African
Americans. Z-angle,
skeletal convexity, lips,
nasomental angle, and
mentolabial sulcus
significant factors in profile
assessment.
Agreement
Maple et
al., 2005
60
Profile
photograph
(color)
3 male, 3
female (n=6);
Age 18-30;
Class I profile
and average
vertical
proportions
Altered AP and
vertical
proportions (9
profiles)
1. Orthodontists, 2.
Oral surgeons, 3.
Laypersons; n=100
VAS Class I preferred; large
deviations from Class I
were less attractive.
Preferences affected by AP
and vertical interactions.
Agreement
9
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Soh et al.,
2005
98
Profile
photograph
(black and
white)
1 Chinese
male, 1
Chinese
female (n=2);
Normal
profile, Class
I incisor,
Class I
skeletal.
Altered AP (7
profiles)
Asian: 1.
Orthodontist), 2.
Oral surgeons, 3.
Dental students, 4.
Laypersons; n=275
Rank order;
Most
influential
feature of
face
Class I or bimaxillary
retrusion Chinese profiles
preferred. Upper lip, lower
lip, and chin most
influential features in a
profile.
Agreement
2-D FACIAL FRONTAL VIEW
Shaw,
1981
95
Frontal
photographs
(black and
white)
2 male, 2
female (n=4)
Attractive vs.
Non-attractive;
Altered
dentofacial
arrangement
1. Childern, 2.
Adults; n=84
VAS More attractive: normal
dental appearance,
associated with positive
traits.
Agreement
Lundstrom
et al.,
1989
55
Annual
frontal
photographs
32 male, 32
female (n=
64); Skeletal
age 12 + 18
Sn-Gn
measurements (3
groups)
1. Orthodontic
faculty, 2. Artists, 3.
Laypersons, 4.
Orthodontic senior
residents, 5.
Orthodontic junior
residents; n=20
1-5 Likert
scale
Horizontal mandibular
growers more attractive
than vertical mandibular
growers. Females more
attractive than males.
Positive correlation
between profile and frontal
view assessments.
Agreement.
Laypersons rated
higher and artists
rater lower.
10
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Matoula
and
Pancherz,
2006
62
Frontal
photographs
Attractive: 5
male, 25
female;
Nonattractive:
21 male, 11
female
(n=62). Age
14-24
Cephalometric
relationship: 1.
Sagittal, 2.
Vertical, 3.
Facial height, 4.
Profile
convexity, 5. Lip
positoin
Dental students
(n=54)
3 point
Likert scale
Frontal view is related only
to a minor degree to lateral
skeletaltofacial indices.
ANB, Wits, profile
convexity, and lips
significant factors in facial
attractiveness.
N/A
2-D FACIAL THREE-QUARTER VIEW
Howells
and Shaw,
1985
41
Three-
quarter
photograph
(color), live
subjects
n=54 Photographs vs.
Live subjects
Laypersons (n=2) VAS Photographs (2-D)
representative of live (3-D)
subjects
N/A
Tatarunaite
et al.,
2005
100
Longitudinal
smiling and
non-smiling
three-quarter
photograph
29 male, 31
female
(n=60). Age
11 and 31
1. Overall, 2.
Cheeks, 3. Chin,
4. Eyes, 5. Hair,
6. Lips, 7. Nose,
8. Skin, 9. Teeth
United Kingdom
laypersons (n=12)
9 point
Likert scale
More attractive: smiles,
youth. Less attractive: age.
People retain their relative
attractiveness throughout
life.
N/A
2-D MULTIPLE
IMAGES
Peck and
Peck,
1970
72
Frontal and
profile
photograph,
and lateral
cephalograph
3 male, 49
female
models
(n=52); Age
21;
Caucasian;
Class I
Cephalometric
relationship
Models have fuller, more
protrusive dentofacial
patterns than cephalometric
standards.
N/A
11
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Bell et al.,
1985
7
Frontal and
profile
photograph
22 males, 58
females
(n=80), age
16-54, 40
surgical, 40
orthodontics-
only
N/A 1. Orthodontists, 2.
Oral surgeons, 3.
Laypersons, 4.
Subjects; n=206
Agreement
between
professional groups
and laypersons; No
agreement between
professional groups
and subjects.
Bittner
and
Pancherz,
1990
12
Frontal and
profile
photographs
(black and
white),
lateral
cephalograph
93 male, 79
female
(n=172), age
12-14
N/A 1. Orthodontists, 2.
Orthodontic
residents; n=7
Sagittal and vertial dental
and skeletal relationships
only partly reflected in the
face. Large OJ reflected in
face.
N/A
Kerr and
O'Donnell,
1990
44
Full-face and
profile
photographic
30 male, 30
female
(n=60);
Mean age
12.4
OJ; Before and
after Treatment
1. Orthodontists, 2.
Dental students, 3.
Art students, 4.
Parents of treated
children, n=16; Age
20-30y (except
parents-older)
5 point
Likert scale
Class I more attractive than
Class II or III; Full face
more attractive than profile
(except for Class I: full face
= profile)
Agreement within
professional groups
and layperson
groups; no
agreement between
professional groups
and laypersons.
Phillips et
al., 1992
78
Frontal and
profile
photographs
13 surgical,
13
orthodontic-
only (n=26)
1. Pre-treatment,
2. Post-treatment
1. Orthodontists, 2.
Orthodontic
residents, 3.
Maxillofacial
surgeons, 4.
Maxillofacial
surgery residents, 5.
Laypersons; n=70
VAS Agreement.
Laypersons more
critical than
orthodontists and
surgeons.
12
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Phillips et
al., 1992
79
Frontal non-
smiling,
frontal
smiling,
profile
photographs
(color)
Class I and
II, surgical
and
orthodontic-
only (n=18)
N/A 1. Orthodontic
residents, 2. Dental
students, 3.
Undergraduate
students; n=104
VAS Attractiveness is affected by
view presented. No one
view is favored.
No agreement.
Orthodontic
residents rated
higher that other
groups.
Proffit et
al., 1992
84
Frontal and
profile
photographs
13 surgical,
13
orthodontic-
only (n=26)
1. Pre-treatment,
2. Post-treatment
1. Orthodontists, 2.
Orthodontic
residents, 3.
Maxillofacial
surgeons, 4.
Maxillofacial
surgery resident;
n=37
VAS More attractive:
orthodontic-only subjects
pre- and post-treatment.
Less attractive: surgical
subjects pre- and post-
treatment. Surgical subjects
had improved esthetics post-
treatment.
Agreement
Tulloch et
al.,
1993
104
Frontal non-
smiling,
frontal
smiling,
profile
photographs
(color) and
lateral
cephalograph
Class I and
II, surgical
and
orthodontic-
only (n=18)
N/A 1. Orthodontic
residents, 2. Dental
students, 3.
Undergraduate
students; n=104
VAS Correlation between AP
measures and facial
attractiveness, especially OJ
(greater OJ = less
attractive).
Agreement
13
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
Phillips et
al., 1995
77
Frontal non-
smiling,
frontal
smiling,
three-
quarter,
profile
photographs
(color)
White
females
(n=19),
Class I
(n=4), Class
II (n=15);
age 20-40
N/A 1. Orthodontists, 2.
Oral surgeon, 3.
Subjects; n=24
Rank order Class I most attractive; Class
II surgical least attractive.
Agreement.
Vargo et
al.,
2003
107
Frontal non-
smiling,
frontal
smiling,
profile
photographs
(color)
20 male, 37
female
(n=57); Age
16-45 years
N/A 1. Orthodontist, 2.
Oral surgeons, 3.
Laypersons; n=32
10 point
Likert scale
N/A Agreement.
Decision to seek
treatment more
correlated to
laypersons' rating.
Knight
and Keith,
2005
49
Frontal,
three-
quarter,
profile
photographs
(color)
30 male, 30
female
(n=60);
Caucasian
N/A 1. Orthodontists, 2.
Laypersons; n=12
Rank order Soft-tissue ANB and
anterior lower facial height
had minimal correlation
with facial attractiveness.
Agreement
Kiekens et
al., 2006
46
Frontal,
three-quarter
smiling,
profile
photographs.
32 boys, 32
girls
(n=64);
White; Age
10-16
Dental (OJ) and
skeletal (ANB,
Sn-GoGn);
Dental esthetics
Laypersons (n=78) VAS Dental esthetics most
important indicator for facial
esthetics.
N/A
14
Table 1. Continued.
Reference Images Patient
Sample
Categories Raters Rating
scale
Facial/profile preference Rater group
agreement
3-D IMAGES
Ferrario et
al., 1997
32
2D frontal
and profile
photographs;
3D infrared
video
system
36 male, 44
female
(n=80); Age
8-9
Attractive vs.
Non-attractive
1. Clinicians
(Orthodontists, oral
surgeons), 2.
Casting agents, 3.
Laypersons; n=21
4 point
Likert scale
Frontal and profile
photographic evaluations
differ. No significant
differentce between
attractive and non-attractive
children by 3-D analysis.
Agreement
Todd et
al.,
2005
103
2D snap
shot, 3dMD
scan (black
and white)
2 male, 2
female
(n=4);
Skeletal
Class II
Altered AP (5
profiles)
1. Orthodontist, 2.
Oral Surgeon, 3.
Public; n=100
Rank order 2-D and 3-D: no consistent
preference for Class I. Sex
was not significant. Older
raters less likely to prefer
Class I faces.
No agreement
15
range of malocclusions has been found even in those we find most attractive,
models.
69
Combining overjet and ANB may be a better predictor.
46
Soft-tissue
measurements used include the E-line and S-line.
20,22,34,62,72
A few studies used
multiple parameters,
12,46,62
while some did not specifically identify the parameters
used.
18,19,52,53,60,83
Additionally, altering components of one image has also been
studied.
14,18,19,22,23,34,52,53,60,83,93,95,98,103
The contribution of vertical parameters, as
measured by Sn-GoGn, to profile preference is unclear;
18,20,23,37,93
however
Lundstrom et al.
55,56
did find an association when using the measurement Sn-Gn.
Some research has shown that cephalometric standards correlate with the
judgments of facial esthetics,
55,56,72,84,89,90,104
while others have found that these
measures do not correlate with the perception of facial esthetics.
12,35,62,63,84,108
Morphologic features that appear to most influence facial esthetics include dental
and skeletal relationships;
44,95,104
soft tissue proportions;
6,20,22,34,53,65
and other
features, such as hairstyle, hair color, skin color, eyes, and
complexion.
8,18,30,32,42,65,66,76,82,101,102,104
Black and white photographic images have
been used to eliminate the influence of color,
12,14,18,19,95,98,103
but this does not
represent reality. Although cephalometric measurements are an important component
of an esthetic evaluation, they are not adequate for a complete evaluation.
72
Presently, orthodontists define esthetics in terms of the profile. This is largely
due to the emphasis that has been put on the lateral cephalometric radiograph and its
analyses, which were introduced in the 1950’s and 1960’s. However, Mackley
57
showed that the profile is not a reliable predictor of appearance. How can a
16
profession diagnose and treat based on information gathered from a view of the face,
the profile, which most individuals don’t even consider or see? The lateral view does
provide essential information about the skeletal morphology,
26,99
but the assessment
of facial attractiveness from the frontal view is more important than from the lateral
view.
44,107
Different views of the face are perceived differently because the skeletal
and dental relationships are more or less visible in different views. Frontal facial
views are rated more attractive for Class II and III patients, while frontal and profile
facial views are equally attractive for Class I patients.
44
Although orthodontists are
trained to assess the profile view, it is less important to the layperson who rarely see
their own profiles.
7,36,55,84,94,108
Preferred profiles: by malocclusion
It is through all of these studies that Mew
65
concluded that the only scientific
means of measuring a subjective quality, such as facial beauty, is through a panel of
judges. Many studies (Table 1) using a panel of judges have concluded that the Class
I profile is preferred.
16,18,19,23,44,52,60,77,89,90,98
The further the deviation is from the
Class I profile, the less attractive it is.
44,60,77
Class II and Class III malocclusions are
rated less facially attractive.
18,19,44
Phillips’ et al.
77
study on the perception of facial
attractiveness of white woman found that skeletal Class I patients were ranked more
attractive, whereas Class II patients who opted for orthognathic surgery were ranked
least attractive.
17
Preferred profiles: by age and sex
Differences between facial preferences for females and males exist.
22,34,53,55,56
Greater facial convexity has been associated with youth and femininity, while
straighter faces suggest age and masculinity.
22,34
Kerr and O’Donnell
44
stated that
age has little effect on the perception of Class I faces. However, their study had an
age difference of only two and a half years, from pre-treatment to post-treatment.
This result cannot be extrapolated to larger age differences. In a longitudinal study
spanning 30 years, facial attractiveness decreased over time, although people tended
to retain their relative levels of attractiveness throughout their life spans.
100
Youthful
facial features, such as large eyes, high cheekbones, and narrow jaws are preferred.
75
Preferred profiles: by ethnicity
The original cephalometric values were based on Caucasian subjects, but
values for different ethnic groups are now recognized. Currently, ethnic differences
in facial attractiveness remain inconclusive. In a study by Tedesco et al.,
102
there was
no difference in the ratings of black and Caucasian subjects; conversely, Hall et al.
39
found that a more convex profile with prominent lips was preferred for African
Americans. The preferred Chinese profile is Class I or bimaxillary retrusive.
98
Preferred profiles: by averageness
Averaged faces, created from a composite of individual faces, are viewed as
more attractive than individual facial photographs.
51
This is evidence to support the
18
idea that averageness is ideal;
28
however, individual attractive faces are rated as more
attractive than an averaged composite face.
69,75
Thus averaged faces are not
optimally attractive.
Professionals vs. laypersons: differing perceptions of esthetic ideals
Research has shown that professionals and laypersons agree on facial
attractiveness.
20,23,27,32,34,39,53,55,56,60,66,77-79,84,89,90,93,95,97,98,104,107
Conversely,
discrepancies have been found between professionals and laypersons in other
research.
7,18,44,72,83,103
However, other research has shown that orthodontists and
patients themselves may have differing perceptions of an esthetic
ideal.
7,18,19,44,83,92,101
It is important to understand how patients view themselves if the
outcome of treatment is to be satisfactory. Features that are perceived as important to
orthodontists may appear insignificant to laypersons or patients.
18
Interestingly, two
studies that investigated the effect of different viewpoints found no agreement
between professionals and laypersons.
79,103
Orthodontists focus on the facial profile
and the lower two-thirds of the face, including the lips and chin; whereas patients
view the face as a whole.
18,19,60
In general, laypersons show greater variation and are
less critical in what is considered attractive.
18,21,44,83
It has been suggested that the
observed differences are related to the variation in training, knowledge, and
experience between the groups.
13,18,36,79,83
With the prevalence of mass media today,
the layperson is becoming more educated in what is perceived as attractive. The
patient’s opinion cannot be disregarded, especially when a patient’s decision to seek
19
treatment correlates more with a self-perceived or layperson’s assessment of
attractiveness.
7,107
Within the professional groups, oral and maxillofacial surgeons tend agree
with orthodontists;
14
however, these professional groups differ significantly when
compared to dental students.
53
Dental students rate in between general dentists and
patients;
13
while dental students nearing the end of their training appeared almost as
critical as the orthodontists.
44
When assessing at raters as one whole group, other trends emerge. The sex of
the rater does not appear to affect facial attractiveness assessments;
13,19,42,102,103,106
although, raters may score the opposite sex slightly higher.
29
However, more women
place appearance as an important feature when compared to men.
29
In addition,
women’s preferences shift during the menstrual cycle, with less masculine male
features being preferred during the low conception phases.
74
The age of the rater may
be a factor, as older raters, specifically over the age of 65, are less likely to prefer the
Class I profile
103
and are less likely to rate youthful faces as attractive.
106
Differing
results have been found when the ethnicity of the rater is evaluated. Foster
34
determined that black and Chinese laypeople prefer fuller female profiles than males
profiles, while white laypeople preferred equal fullness. In studies by Tedesco et
al.
102
and Hall et al
39
, there was no difference in the ratings by African American
and Caucasian raters. Conversely, Martin
61
found that white and black American
men preferred white features in female faces, whereas black African men preferred
black features in females. Recently, both Langlois et al.
50
and Naini et al.
71
stated
20
that there is a similar esthetic judgment across various cultural backgrounds. Finally,
raters with lower socioeconomic class, as determined by occupation, scored lower
ratings for facial attractiveness.
41
The future: 3-D technology
Historically, 2-D images have been used to assess facial esthetics. Although
there is a good correlation between live subjects and photographs,
41
2-D imaging has
limitations in the evaluation of 3-D structures.
9
Some of these limitations include a
lack of perspective, effects of projection and superimposition, artifacts, voids, and a
lack of motion.
59
Three-D technologies have seen rapid growth in the past several decades and
their applications in orthodontics are growing. Traditional 2-D records, photographs
and radiographs, are being supplemented or replaced with 3-D records, CBCT and
photo-realistic surface image. Three-D records are not only of interest to
orthodontists, but to oral and maxillofacial or plastic surgeons. Research is underway
to create treatment and surgical simulation software that will allow for treatment and
surgical planning and outcome visualization.
64
The Dolphin 3D software is able to
mesh a 3-D surface image with a CBCT volume. This easy to use software will assist
with patient education and will facilitate moving surgical predictions from 2-D to 3-
D.
Non-invasive, low-radiation craniofacial imaging has been achieved through
cone-beam computed tomography (CBCT). Radiation exposure has been reduced to
21
40-50 µSv or levels similar to a conventional dental radiographic exam.
54,58
A great
deal of research has been done or is currently being completed to create 3-D imaging
standards. Because longitudinal data is limited, research will continue until these
voids can be filled. As the cost and the radiation exposure decrease, and the speed of
imaging and the resolution increase, its uses and applications have expanded.
Similar to the cephalometer’s introduction in the 1950’s or to any new
technology, 3-D imaging has not been fully accepted by the clinical orthodontist.
Standardization of this technology will aid the clinician in accepting this complex yet
complete technology. Initially, 3-D images were reduced to 2-D slice information,
which produced radiographic images similar to the 2-D images the clinicians were
accustomed to viewing. Not only does this reduce the benefits of 3-D imaging, but it
introduces error. The information that can be obtained from 3-D imaging includes
actual measurements, spatial images, rotating images, “peeling” outer structures to
view inner structures, changing density to view different structures, and viewing
asymmetries.
Prior to the introduction of CBCT, methods to surface image the face in 3-D
included videos,
33,38
laser scanners,
68
and stereophotogrammetry.
15,86,87
Although
CBCT does produce an image of the facial soft-tissue, CT scans are used primarily to
identify hard-tissue structures, such as bone. The facial soft-tissue image produced
may be distorted due to a supine imaging position and is devoid of color. No
research has compared the facial soft-tissue image produced by the different
methods.
22
Stereophotogrammetry was developed from technology used to plot terrain
maps from aerial photographs and applied to create contour maps of the face.
15
The
scanner captures multiple views that are processed by stereo-matching algorithms to
produce a 3-D image. Two types of stereophtogrammetric surface imaging
techniques are currently being used. Active stereophotogrammetry (e.g., 3dMD (3Q
Technologies, Atlanta, GA)) uses a grid to register the images. The passive system
(e.g., Di3D (Dolphin Imaging, Chatsworth, CA) uses imperfections or distortions
within the image for alignment. Software to integrate the CBCT and surface image
has been developed, allowing for an accurate 3-D image that correlates the soft-
tissue with the supporting skeleton. Unfortunately, the eyes and ears are not
accurately imaged due to tissue reflectance and undercuts in these areas. Also,
interferences with the hair and eyebrows prevents accurate imaging of these areas.
59
Comparison of 2-D and 3-D techniques
With greater accessibility of 3-D technologies, orthodontic evaluations are
beginning to be made in 3-D. Currently, facial esthetic research refers to 2-D images.
Because the perception of attractiveness is affected by the view presented and no one
view is consistently preferred, showing multiple 2-D views simultaneously allows
for a more complete assessment of facial attractiveness.
79
Three-D images are
expected to be superior as they provide all possible views of a patient in one record.
Research on 3-D facial esthetics is limited.
23
Two studies have compared 2-D and 3-D facial images. Ferrario et al.
32
had a
panel separate attractive and non-attractive children based on 2-D frontal and profile
photographs. No differences were found between these two groups when facial
morphometric characteristics from a 3-D infrared video system were compared.
Todd et al.
103
directly compared 2-D and 3-D images. The variation in the results
was too great to conclude that a difference exists.
The aims of this study were to determine whether viewing facial images in 2-
D and 3-D has any effect on the perception of facial attractiveness and to determine
if the preferred facial relationship chosen by professionals and laypersons in 2-D and
3-D was Class I. Conclusive evidence from a comparison of 2-D and 3-D facial
images will dictate if the historical research on facial esthetic in 2-D can be applied
to 3-D.
24
Chapter 3: Materials and Methods
The sample was comprised of 20 subjects (14 females and 6 males) from
archival data at a Los Angeles private orthodontic practice. Three-D facial surface
imaging (Di3D) (Dolphin Imaging, Chatsworth, CA) and cone-beam computer
tomography volumes (CBCT) (iCAT) (Imaging Sciences International, Hatfield, PA)
were available for all subjects as part of the orthodontic records.
The following were inclusion criteria for the subjects:
(1) presented for comprehensive orthodontic treatment and have archived images
(2) age 18 to 35
(3) sagittal cephalometric measurement (ANB) within 1 SD of normal for facial
profile type Class I and 2 SD for Class II and III (Table 1)
(4) vertical cephalometric measurements (Sn-Go-Gn, FMA) within 1 SD of normal
for all facial profile types
(5) clearly defined Angle molar classification
(6) acceptable image quality
(7) absence of craniofacial anomalies
(8) absence of distracting facial characteristics (eg. facial hair).
Table 2. Cephalometric values classifying subjects into three facial profile types (I,
II, III) based on Steiner
99
norms
Class I Class II Class III
Mean (SD) Mean (SD) Mean (SD)
ANB (º) 2 (2) 4 (2) 0 (2)
Wits (mm) -1 (1) 1 (1) -2 (1)
Dolphin Imaging 10.1 with Dolphin 3D (Dolphin Imaging) was used to
reconstruct and trace a lateral cephalograph in 2-D for each subject. The sample was
classified into three facial profile types: Class I (n = 9), II (n = 7), and III (n = 4).
The classification was based on the relationship of the maxilla and mandible to the
cranial base (ANB) and to the occlusal plane (Wits) (Table 1). The surface image
and the CBCT volume were aligned and had Frankfort horizontal parallel to the
horizon.
To maintain image characteristics (eg. lighting, color, resolution,
magnification, posture), the 2-D images (frontal and profile) were reconstructed from
snap-shot views of the 3-D surface image. The frontal and profile 2-D images were
placed side by side (Figure 1). A video clip was constructed from the 3-D surface
Figure 1. 2-D images (frontal and profile)
image within Dolphin 3D using Windows Media Player version 10.0 (Microsoft,
Tacoma, WA). The video clip moved the image from right profile, to frontal, to left
profile, to frontal, to submentovertex, to chin down, and to frontal (Figure 2).
25
Figure 2. 3D surface image video clip
The 2-D and 3-D surface images were randomly ordered in a Microsoft
PowerPoint 2003 (Redmond, WA) presentation (Figure 3). Five versions of the
presentations with different random image orders were used in order to limit
temporal biases.
Align CBCT + 3-D Surface Image
with FH parallel to horizon
n = 60
Create and trace 2-D Cephalometric
Radiographs
3-D Surface Image Video
n = 20
2-D Snapshot
Photographs from 3-D
Surface Image
(1) Profile
(2) Frontal
n = 20
PowerPoint Presentation
(n = 5)
-Subjects in random order
Excluded Subjects
n = 40
Figure 3. A flow chart of the image creation process
26
27
The raters were recruited to participate through an email request (Appendix
1) or verbally. The group consisted of 86 people, 33 orthodontists and orthodontic
residents, 26 dental students, and 27 laypeople. The following were inclusion criteria
for the raters:
(1) orthodontists, including orthodontic residents
a. orthodontists who are USC, School of Dentistry full-time or part-time
faculty
b. orthodontic residents at USC, School of Dentistry
(2) dental students at USC, School of Dentistry
(3) laypeople
a. employees at USC, School of Dentistry
b. friends
(4) signed the agreement to participate (Appendix 2) in the study and to protect
the confidentiality of the subject.
Any rater who knew the subjects presented as stimulus faces were excluded
from participating in the study.
The raters were not be given any specific information about the faces they
saw, save that they were to assess the facial attractiveness of each image and that
there was no correct answer.
28
Study 1 – Rating
The visual analog scale (VAS) was used for recording the scores. The 100
mm scale will be anchored on the left with “least attractive” and on the right with
“most attractive.” The raters were instructed on what they were to do and what to
expect to see (Appendix 2). A series of two introductory images (one 2-D image and
one 3-D video) was shown to accustom the raters to the procedure. The 2-D images
were shown for 12 seconds and the 3-D images were shown for 25 seconds. A blank
screen was shown for a period of 4 seconds following each image to allow time for
scores to be recorded.
The individual ratings were measured in millimeters with electronic calipers
(Pro-Max, Fowler-NSK, Newton, MA) calibrated to the nearest one hundredth of a
millimeter (0.01mm).
Study 2 – Ranking
A subset from each of the three profile types was randomly chosen (n = 5
males, n = 6 females) to have images ranked in order of preference. The male and
female images were ranked separately. All five male 2-D images were viewed at the
same time. The raters were asked to rank the images in order of preference from 1
(most attractive) to 6 (least attractive). The raters were given 45 seconds to rank the
2-D images. The raters were not permitted to assign the same ranking to more than
one image. The ranking were repeated for the female group. Rankings were done in
the same manner with the 3-D surface images and were viewed for 60 seconds.
29
Statistical Method
The statistical analysis was performed using Microsoft Excel 2003
(Redmond, WA), SPSS 14.0 (Chicago, IL), and SAS 9.1 (Cary, NC). Descriptive
statistics (means and standard deviations) were computed for each variable.
Student’s t-tests for independent samples and Levene’s test for equality of variance
were used to compare the age and sex of the subjects, as well as the age and sex of
the raters. The subjects’ ages and cephalometric values were analyzed by facial
profile type using one-way ANOVA. Post-hoc testing was done with the Tukey HSD
method. The level of significance was 0.05.
Independent t-tests were used to compare two and three dimension scores.
One-way ANOVA was used to compare scores across subject groups and rater
groups independently, within each dimension. Because the scores varied by subject
group and by rater group, multivariate analyses were conducted using a generalized
linear model (PROC GLM) to determine whether the scores were significantly
different by dimension across subject and rater groups, adjusting for age, sex, and
ethnicity and mutually adjusting for rater and subject groups. Associations by
dimension were further analyzed within subject and rater groups, respectively, using
a generalized linear model with Bonferroni’s correction. All p-values are two-sided
with a significance level of 0.05.
The method error was assessed using the intra-class correlation coefficient
(ICC) with a 95% confidence internal. All lateral cephalometric radiographs were
evaluated for inclusion criteria and were traced by one examiner. Duplicate
30
cephalometric measurements of one subject taken by the same examiner two weeks
apart were analyzed. Intra-rater variance was assessed by analyzing the difference
between duplicate measurements taken two weeks apart on all subjects by one rater.
31
Chapter 4: Results
Subjects
The demographic data of the subjects (Table 1) show that the number in each
group, by facial profile type, was not evenly distributed (20 overall, 9 Class I, 7
Class II, 4 Class III). More females than males made up the overall, Class I, and II
groups, while the Class III group had an equal number of females and males. The
mean age by facial profile type was not significantly different (p = 0.190) between
all three groups with an overall mean (SD) of 26.4 (5.9) years. The cephalometric
tracing reliability was 0.99. There were significant differences in the sagittal
cephalometric measurement, ANB angle, and Wits-appraisal (p < 0.0001), between
Table 3. Subject demographic data
Characteristic Overall
Class I
(n = 9)
Class II
(n = 7)
Class III
(n = 4) P-value
Sex
Male 6 3 1 2
Female 14 6 6 2
Age (years)
Means (SD) 26.4 (5.8) 25.0 (6.2) 29.6 (5.5) 24.0 (4.0) 0.190
Range 18-35 18-32 21-35 19-29
Cephalometric
values (degrees)
ANB (SD) 2.2 (3.6) 1.3 (0.9) 6.3 (0.4) -3.0 (1.6) <0.0001
Wits (SD) -0.9 (5.5) -0.6 (0.8) 4.0 (2.0) -10.0 (3.2) <0.0001
Sn-GoGn (SD) 32.8 (3.5) 31.4 (1.8) 35.0 (3.8) 32.3 (4.6) 0.103
FMA (SD) 24.2 (3.1) 22.1 (1.3) 25.9 (2.7) 26.3 (4.1) 0.010
32
the subject groups. There was no significant difference between the groups in the
vertical measurement Sn-GoGn, however there was a significant difference in the
FMA angle (p = 0.010*).
Raters
The total number of raters was 86 (27 laypersons, 33 orthodontists, 26 dental
students) (Table 2). There were more male raters in all groups; except for the dental
students, where there were more females. There was a wide variation in age for each
group of raters. The laypersons and orthodontists had similar age means and ranges.
The dental students were significantly younger (p <0.001) and had a smaller age
range, when compared with the other two groups. Five ethnic groups were
Table 4. Rater demographic data
Characteristic
Overall
(n = 86)
Laypersons
(n = 27)
Orthodontist
(n = 33)
Dental
Students
(n = 26) P-value
Sex
Male 47 15 21 11
Female 39 12 12 15
Age (years)
Means (SD) 37.2 (14.7) 40.8 (15.0) 42.2 (16.1) 27.0 (3.5) <0.0001
Range 18-70 18-70 26-70 21-35
Ethnicity
Caucasian 57 22 10 16
Hispanic 4 4 0 0
Asian 19 1 11 7
Middle Eastern 5 0 2 3
East Indian 1 0 1 0
33
represented amongst the raters. Two-thirds of the raters were Caucasian, 22% were
Asian, and the remaining 12% were of Hispanic, Middle Eastern, or East Indian
ethnicity. The intra-rater reliability was 0.79.
Rating: Comparison of 2-D and 3-D images
The rating measurement reliability was 0.99. The rating scores ranged from 0
to 96mm. Mean values were computed within each subject and rater group, and
comparisons were made between the two dimensions (Table 4). The mean scores in
3-D were significantly higher, than those in 2-D (p < 0.0001). The same significant
differences were found across all subject and rater groups. The direction of the
association between 2-D and 3-D scores did not vary by subject or rater groups.
Within the subject groups, the mean scores were significantly different for
each subject group with Class I subjects highest, followed by Class II subjects, and
finally Class III subjects. The difference between dimension and subject group mean
scores was greatest for Class I (p < 0.001) and lowest for Class III (p < 0.05). Within
the rater groups, the mean scores were significantly different for each rater group
with the scores given by the orthodontists highest, followed by the laypeople, and
with the lowest scores given by the dental students. The difference between
dimension and rater group mean score was greatest for the orthodontists (p = 0.001)
and least for dental students (p < 0.05). The association was weakest for Class III and
dental students. Within these two lowest rated groups, the difference between 2-D
34
and 3-D is less compared to the differences observed in the other subject and rater
groups (p < 0.05).
Table 5. Comparison of mean score in overall, and by subject and rater groups
Two dimension
scores
Three dimension
scores
Mean (SD) Mean (SD)
P-value
Overall 43.6 (20.4) 46.7 (21.3) <0.0001
By subject group
Class I 51.1 (19.9) 54.8 (20.4) 0.0004
Class II 42.2 (18.4) 45.0 (19.3) 0.01
Class III 29.1 (16.0) 31.6 (17.2) 0.04
P-value <0.0001 <0.0001
By rater group
Laypersons 41.0 (19.3) 43.8 (19.9) 0.02
Orthodontists 48.9 (20.2) 52.5 (20.5) 0.001
Dental students 39.6 (20.5) 42.3 (22.0) 0.04
P-value <0.0001 <0.0001
When adjusted for age, sex, and ethnicity between subject and rater groups
(Table 5), there was no significant difference between the dimension and scores by
subject group (p = 0.74) or by rater group (p = 0.78). There was a significant
difference between dimensions for each subject and rater group. The direction of the
association between dimension and scores is the same for all groups. The same
subject and rater group order of preference was found as in Table 4, with Class I and
orthodontists rated highest.
Within each dimension, scores varied significantly among rater groups in the
overall sample and by each subject group (Table 6). The highest scores were
obtained for the 3-D images, for the Class I subjects, and by the orthodontists. The
35
Table 6. Multivariate association between dimensionality of image and score in
overall sample and by subject and rater groups
Two dimension Three dimension
LS Mean (SE) LS Mean (SE)
P-value
Overall 37.4 (0.81) 40.5 (0.81) <0.0001
By subject group
Class I 48.0 (1.23) 51.6 (1.23) <0.0001
Class II 38.3 (1.35) 41.0 (1.35) 0.008
Class III 26.6 (1.61) 29.2 (1.61) 0.04
P-value <0.0001 <0.0001
By rater group
Laypersons 34.5 (1.24) 37.3 (1.24) 0.01
Orthodontists 47.0 (1.07) 50.7 (1.07) 0.0001
Dental students 32.2 (0.95) 34.9 (0.94) 0.02
P-value <0.0001 <0.0001
Bonferroni correction determined that the orthodontists scored significantly higher in
the overall sample and in all three subject groups (all p-values < 0.001). Scores from
the laypersons and dental students were not significantly different in the overall
sample or within each subject group.
Within each dimension, scores varied significantly between subject groups in
the overall sample and by each rater group (Table 7). The highest scores were
obtained for the 3-D images, for the Class I subjects, and by the orthodontists. The
Bonferroni correction determined that the Class I subjects scored significantly higher
in the overall sample and in all three rater groups (all p-values < 0.001). Between
subject groups, scores differed significant for all three groups.
36
Table 7. Association among rater group and score by dimension in overall sample
and by subject groups
Laypersons Orthodontists Dental students
LS Mean (SE) LS Mean (SE) LS Mean (SE)
P-value
Overall
2-D 33.6 (1.26) 43.0 (1.11) 34.4 (1.37) <0.0001
3-D 37.3 (1.30) 47.6 (1.15) 38.1 (1.42) <0.0001
By subject
group
Class I
2-D 41.5 (1.94) 54.1 (1.71) 44.2 (2.11) <0.0001
3-D 47.0 (1.98) 61.5 (1.73) 50.4 (1.98) <0.0001
Class II
2-D 35.6 (2.11) 43.9 (1.85) 36.0 (2.30) <0.0001
3-D 38.1 (2.21) 46.6 (1.94) 37.8 (2.40) <0.0001
Class III
2-D 26.2 (2.49) 30.6 (2.19) 23.9 (2.71) 0.01
3-D 28.7 (2.65) 32.5 (2.34) 25.5 (2.88) 0.02
Table 8. Association among subject group and score by dimension in overall sample
and by rater groups
Class I Class II Class III
LS Mean (SE) LS Mean (SE) LS Mean (SE)
P-value
Overall
2-D 47.4 (1.13) 38.4 (1.18) 25.3 (1.34) <0.0001
3-D 52.0 (1.17) 41.2 (1.22) 28.8 (1.38) <0.0001
By rater group
Laypersons
2-D 43.7 (1.74) 36.7 (1.84) 23.7 (2.15) <0.0001
3-D 45.9 (1.81) 38.8 (1.92) 26.7 (2.24) <0.0001
Orthodontists
2-D 58.3 (1.55) 46.4 (1.64) 32.4 (1.91) <0.0001
3-D 65.6 (1.53) 52.8 (1.62) 37.6 (1.88) <0.0001
Dental student
2-D 41.5 (1.32) 34.3 (1.48) 22.1 (1.87) <0.0001
3-D 44.4 (1.40) 35.5 (1.55) 23.2 (1.98) <0.0001
Ranking: Comparison of 2-D and 3-D images
Frequencies were calculated within subject group by dimension for the
overall sample. Males were not calculated due to the unequal distribution by subject
group. In 2-D, female Class I was rated highest, followed by Class II, and Class III
Class I female
0
10
20
30
40
50
12 345 6
Rank
Percentage (%)
2D
3D
Figure 4. Rank frequencies of Class I females by dimension
Class II female
0
10
20
30
40
50
12 345 6
Rank
Percentage (%)
2D
3D
Figure 5. Rank frequencies of Class II females by dimension
37
Class III female
0
10
20
30
40
50
12 345 6
Rank
Percentage (%)
2D
3D
Figure 6. Rank frequencies of Class III females by dimension
(Figure 4, 5, 6). In 3-D, the opposite trend was found with female Class III rated
highest, followed by Class II, and Class I (Figure 4, 5, 6).
Frequencies were calculated within rater groups by dimension and between
subject groups. Males were not calculated due to the unequal distribution by subject
group. In 2-D, all three rater groups preferred Class I, followed by Class II, and Class
III (Figure 7, 8, 9). In 3-D, nearly the opposite trend was found with female Class II
and Class III subjects rated highest and Class I subjects rated lowest (Figure 10, 11,
12).
38
2D female facial profile preferences of
Laypersons
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 7. Rank frequencies by laypersons by subject group in 2-D
2D female facial profile preferences of
Orthodontists
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 8. Rank frequencies of orthodontist by subject group in 2-D
39
2D female facial profile preferences of
Dental students
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 9. Rank frequencies of dental student by subject group in 2-D
3D female facial preferences of
Laypersons
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 10. Rank frequencies by laypersons by subject group in 3-D
40
3D female facial profile preferences of
Orthodontists
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 11. Rank frequencies of orthodontist by subject group in 3-D
3D female facial preferences of
Dental students
0
10
20
30
40
50
12 3 45 6
Rank
Percentage (%)
Class I
Class II
Class III
Figure 12. Rank frequencies of dental student by subject group in 3-D
41
42
Chapter 5: Discussion
Technological advances are changing the face of orthodontics. The primary
objective of this study was to compare the facial attractiveness preferences of 2-D
and 3-D images. Secondarily, we tested facial profile preferences, and compared the
preferences of orthodontist, dental students, and laypersons. This study resulted in
significant findings that will aid those developing, researching, and using these new
3-D technologies.
Reliability
Findings from our study on intra-rater reliability (ICC = 0.79) with the VAS
agree with other reports in the literature.
41,60,77
The VAS proved to be a practical
method for assessing facial attractiveness. The continuous variable from 0 to 100
millimeters prevented rater biases towards fixed points, is more sensitive, and allows
greater freedom in the data analysis.
41
However, evaluating facial esthetics, a subjective measure with an objective
tool has inherent faults. In this study, the raters were not given any information on
how to rate, except to use their own personal judgments. As a result, it is difficult to
know how the scale and its anchor points were interpreted by the different raters.
Most importantly, it is uncertain how to interpret these differences into a clinical
significance.
43
Subjects
The subjects all presented themselves for orthodontic records at a Los
Angeles area orthodontic practice. Although subject ethnicity was not recorded, most
subjects were Caucasian. As a result, the findings from this study cannot be
extrapolated to the general population of multiple ethnicities from diverse areas.
Although the subjects were not evenly distributed by sub-group or sex, they were
representative of North American orthodontic practices which include fewer Class
III and more female patients.
At this time, no 3-D cephalometric analysis has been standardized; however a
great deal of research is taking place in this area. In this study, 2-D lateral
cephalometric radiographs were reconstructed from the 3-D volume and evaluated
based on traditional 2-D measurements. The inherent problem with this is that
conventional 2-D radiographs are derived from a perspective projection with built in
magnification, whereas with 3-D CBCT volumes, the projectional magnification is
corrected to create an orthogonal image with a 1:1 ratio. Adams et al.,
1
found 3-D
imaging measurements to be more accurate than traditional 2-D cephalometric
measurements when compared to actual measurement. Chidiac et al.,
17
noted that 2-
D cephalometric and 3-D CBCT volumes were similar in angular, but not linear
measurements. Angular and linear measurements were used in this study. No
significant difference was found between subject groups and the Sn-GoGn
measurement, but a small significant difference in the FMA measurement was found.
44
This may be explained by the greater error in locating the landmarks associated with
Frankfort horizontal, porion and orbitale.
Raters
The raters responded to a request to participate in the study. Although the
study was conducted in the Los Angeles area, the raters were from all across North
America. There were more males than females the overall, the orthdontists’ and
laypersons’ group, but the opposite was true for the dental students. The age ranges
for all groups were significantly different, but were inclusive of the subjects age
range, but with a much higher limit. Two-thirds of the raters were Caucasian, which
matched the mainly Caucasian subject group.
Rating: Comparison of 2-D and 3-D images
When the 2-D and 3-D images were compared, a clear difference existed.
Three-D images were viewed as more facially attractive than 2-D images. The 3-D
perspective and lack of projection or superimposition obtained when the face, a 3-D
structure, is imaged in 3-D as compared to 2-D, may have resulted in our findings.
Despite being statistically significant, the difference between 2-D and 3-D images
was only approximately three millimeters. It is unclear how large a difference on the
VAS is necessary to be clinically relevant.
Subjects were analyzed by facial profile type and a similar significant
outcome was achieved. Three-D images were rated higher for all three profiles. The
45
association was weakest for the Class III profile. The preferred profile was Class I.
These findings, for a general preference of a Class I profile, are in concordance with
other research.
16,18,19,23,44,52,60,77,89,90,98
Although Todd’s
103
research comparing 2-D
and 3-D images was inconclusive, our larger sample size and different testing
methods enabled a decisive results.
Despite differing opinions amongst the groups, overall, the rater groups
followed the same trends observed for the other variables. Although the raters did
not agree on the same facial attractiveness score, the trends between the groups were
similar. Three-D images were rated significantly higher by all three groups. The
orthodontists rated significantly higher for all subject groups. The laypeople and the
dental students rated comparably. These findings concurred with studies where there
was no agreement between the assessors.
7,18,44,72,83,103
On the other hand, other
studies have found agreement between the assessors.
20,23,27,32,34,39,53,55,56,60,66,77-
79,84,89,90,93,95,97,98,104,107
When there is a discrepancy between the assessors, the
laypeople are typically less critical than other groups,
55,56,78
a finding that differs
from ours. Age, sex, and ethnicity of the rater did not alter the results. The
differences between the raters could be explained by education, training, and societal
variables. Therefore, because of the potential for differences between the esthetic
ideals perceived by patients’ and orthodontists’, communication and an attempt to
understand the patients’ expectations is essential if the outcome is to be satisfactory.
46
Ranking: Comparison of 2-D and 3-D images
When 2-D and 3-D images were compared for the overall sample, a clear
trend emerged. In 2-D, there was a strong preference for Class I as the preferred
facial profile and for Class III as the least attractive facial profile. This agrees with
the other method tested in this study, rating using the VAS, and other
research.
16,18,19,23,44,52,60,77,89,90,98
In 3-D the opposite trend was noted, with Class III as
the most preferred and Class I as least preferred. This disagrees with the other
method tested in this study and questions the validity of this method and of 3-D
imaging.
When the rater groups were compared, the 2-D trend was as expected;
however, the 3-D was different from expected. In 2-D, the order from most to least
preferred facial profile was Class I, II, and III. In 3-D, Class II and III were most
preferred, while Class I was least preferred. This difference between 2-D and 3-D
may have been due in part to the methodology, with insufficient time allotted to view
the 3-D images, or to an actual difference in the perception of the 3-D images.
Limitations and future research
In order to minimize variables such as head posture, lighting, color,
resolution, and magnification, the 2-D images, in this study, were reconstructed from
snap-shot views of the 3-D surface image. A subsequent study comparing 2-D
photographs and 2-D snap-shots from 3-D surface images would validate our study.
47
The author’s image was used as an introductory test image and numerous
raters commented that there appeared to be a difference between the 3-D surface
image and reality, with the 3-D image being less attractive. When gathering the data,
the author had noticed a similar trend between the subjects’ 3-D surface image and
2-D photographs. This may be due to the inanimate nature of the 3-D image, an
atypical state. Faces are dynamic structures that appear different when animated.
37,91
Three-D surface imaging is currently being refined. With the Di3D system,
there are distortions in the eyes due to a reflection on the retina, and the ear or chin
areas due to the large changes in contour. There are also problem in the imaging of
the hair, eyebrows, and eyelashes. These distortions may have affected our results. In
addition, these areas may have affected the assessment of the laypeople to a greater
extent. Laypeople view the face as a whole, while orthodontists focus on the profile.
In orthodontics, the eyes, ears, and hair are unaffected by treatment; therefore these
should not limit the application of 3-D imaging. A study comparing the eyes in the
two formats would determine whether the distortion in the 3-D eye would affect the
perception of facial attractiveness between the 2-D and 3-D images.
Despite the many advantages of using 3-D images for rating facial
attractiveness, there are limitations when compared to 2-D images. Most importantly,
clinicians are not accustomed to seeing 3-D records. Two-D photographs and
radiographs have been part of the tangible components of orthodontic records for
decades. Three-D technology is available and 3-D orthodontic records will be used
not only for research purposes, but by the clinical orthodontist. The results of this
48
study indicate a constant difference between the esthetic preferences garnered from
the two image formats. Further research is necessary to validate facial preferences in
3-D.
Although statistically different results were obtained between 2-D and 3-D,
this may not represent a clinically difference. At this time, we cannot conclude if the
historical facial esthetic research based on 2-D can be applied directly to 3-D images.
Differences existed in the actual scores assessed, but the overall trend was the same
for all rater groups: the Class I facial profile was preferred. More research needs to
be conducted to answer the new questions posed, but this study showed that 3-D
imaging, using the Di3D system, is a valid tool for assessing facial attractiveness.
Esthetics will always be a very complex and subjective measure, but new 3-D
technologies are making the tools used to make assessments more like reality and
hopefully more accurate.
49
Chapter 6: Conclusion
Facial surface imaging, a new 3-D technology, is available for clinical
orthodontic use, but limited research is available comparing it to traditional tools.
This information is invaluable to companies trying to improve their tools, but also to
clinicians already using or considering using this technology. The results of this
study, although not generalizable to the population, indicate significant differences in
the perception of facial attractiveness by dimension, facial profile type, and rater.
(1) Three-D images are perceived more facially attractive than 2-D images.
(2) Class I was the preferred facial profile type in 2-D and 3-D. Class II was the
second most preferred facial profile and Class III was the least preferred facial
profile type in both dimensions. There was a concordance between professionals
and laypersons in the preferred facial profile types.
(3) Orthodontists rated higher than dental students and laypersons.
50
Bibliography
1. Adams GL, Gansky SA, Miller AJ, Harrell WE, Jr., Hatcher DC. Comparison
between traditional 2-dimensional cephalometry and a 3-dimensional approach on
human dry skulls. Am J Orthod Dentofacial Orthop 2004;126:397-409.
2. Adams GR. The effects of physical attractiveness on the socialization process.
Psychological aspects of facial form. Ann Arbor: University of Michigan, Center for
Human Growth and Development; 1980: p. 25-47.
3. Albino JE, Cunat JJ, Fox RN, Lewis EA, Slakter MJ, Tedesco LA. Variables
discriminating individuals who seek orthodontic treatment. J Dent Res
1981;60:1661-1667.
4. Angle EH. Treatment of malocclusion of the teeth and fractures of the maxillae.
SSW Den Mfg Co; 1900.
5. Baldwin D. Appearance and aesthetics in oral health. Community Dent Oral
Epidemiol 1980;8:244-256.
6. Bass NM. The aesthetic analysis of the face. Eur J Orthod 1991;13:343-350.
7. Bell R, Kiyak HS, Joondeph DR, McNeill RW, Wallen TR. Perceptions of facial
profile and their influence on the decision to undergo orthognathic surgery. Am J
Orthod Dentofacial Orthop 1985;88:323-332.
8. Bergman RT. Cephalometric soft tissue facial analysis. Am J Orthod Dentofac
Orthop 1999;116:373-389.
9. Berkowitz S. A multicenter retrospective 3D study of serial complete unilateral
cleft lip and palate and complete bilateral cleft lip and palate casts to evaluate
treatment: part 1—the participating institutions and research aims. Cleft Palate
Craniofac J 1999;36:413-424.
10. Birkeland K, Boe OE, Wisth PJ. Orthodontic concen among 11-year-old children
and their parents compared with orthodontic treatment need assessed by index of
orthodontic treatment need. Am J Orthod Dentofac Orthop 1996;110:197-205.
11. Bishara SE, Jakobsen JR. Profile changes in patients treated with and without
extractions: assessments by lay people. Am J Orthod Dentofac Orthop
1997;112:639-644.
12. Bittner C, Pancherz H. Facial morphology and malocclusions. Am J Orthod
Dentofacial Orthop 1990;97:308-315.
51
13. Brisman AS. Esthetics: a comparison of dentists’ and patients’ concepts. JADA
1980;100:345-352.
14. Burcal RG, Laskin DM, Sperry TP. Recognition of profile change after simulated
orthognathic surgery. J Oral Maxillofac Surg 1987;45:666-670.
15. Burke PH, Banks P, Beard LF, Tee JE, Hughes C. Stereophotographic
measurement of change in facial soft tissue morhpology following surgery. Br J Orl
Surg 1983;21:237-245.
16. Burstone CJ. The integumental contour and extension patterns. Angle Orthod
1959;29:93-104.
17. Chidiac JJ, Shofer FS, Al-Kutoubi A, Laster LL, Ghafari J. Comparison of CT
sonograms and cephalometric radiographs in craniofacial imaging. Orthod
Craniofacial Res 2001;5:104-113.
18. Cochrane SM, Cunningham SJ, Hunt NP. A comparison of the perception of
facial profile by the general public and 3 groups of clinicians. Int J Adult Orthod
Orthognath Surg 1999;14:291-295.
19. Cochrane SM, Cunningham SJ, Hunt NP. Perceptions of facial appearance by
orthodontists and the general public. J Clin Orthod 1997;31:164-168.
20. Cox NH, van der Linden FP. Facial harmony. Am J Orthod 1971;60:175-183.
21. Cunningham SJ. The psychology of facial appearance. Dent Update
1999;26:438-443.
22. Czarnecki ST, Nanda RS, Currier GF. Perceptions of a balanced facial profile.
Am J Orthod Dentofacial Orthop 1993;104:180-187.
23. De Smit A, Dermaut L. Soft-tissue profile preference. Am J Orthod 1984;86:67-
73.
24. Dierkes JM. The beauty of the face: an orthodontic perspective. JADA
1987;Dec:89E-95E.
25. Dion K, Berscheid E, Walster E. What is beautiful is good. J Pers Soc Psychol
1972;24:285-290.
26. Downs WB. Analysis of the dentofacial profile. Am J Orthod 1956;26:191-211.
27. Dunlevy HA, White Jr RP, Turvey TA. Professional and lay judgment of facial
esthetic changes following orthognathic surgery. Int J Adult Orthod Orthognath Surg
1987;2:151-158.
52
28. Elder RJ. Background considerations to facial aesthetics. Br J Orthod
2001;28:159-168.
29. Eli I, Bar-Tal Y, Kostovetzki I. At first glance: social meanings of dental
appearance. J Pubilc Health Dent 2001;61:150-154.
30. Enlow DH. Facial Growth. Philadelphia: Saunders; 1990.
31. Farkas LG, Cheung G. Facial asymmetry in healthy North American Caucasians.
An anthropometrical study. Angle Orthod 1981;51:70-77.
32. Ferrario VF, Sforza C, Poggio CE, Colombo A, Tartaglia G. The relationship
between facial 3-D morphometry and the perception of attractiveness in children. Int
J Adult Orthod Orthognath Surg 1997;12:145-152.
33. Ferrario VF, Sforza C, Poggio CE, Tartaglia G. Distance from symmetry: a
three-dimensional evaluation of facial asymmetry. J Oral Maxillofac Surg
1994;52:1126-1132.
34. Foster EJ. Profile preferences among diversified groups. Angle Orthod
1973;43:34-40.
35. Giddon DB. Orthodontic applications of psychological and perceptual studies of
facial esthetics. Semin Orthod 1995;1:82-93.
36. Giddon DB. Through the looking glasses of physician, dentists and patients.
Perspectives Biol Med 1983;26:451-458.
37. Giddon DB, Sconzo R, Kinchen JA, Evans CA. Quantitative comparison of
computerized discrete and animated profile preferences. Angle Orthod 1996;66:441-
448.
38. Gross MM, Trotman C, Moffatt KS. A comparison of three-dimensional and
two-dimensional analyses of facial motion. Angle Orthod 1996;66:189–194.
39. Hall D, Taylor RW, Jacobson A, Sadowsky PL, Bartolucci A. The perception of
optimal profile in African American versus white Americans as assessed by
orthodontists and the lay public. Am J Orthod Dentofac Orthop 2000;118:514-525.
40. Hellmann M. Some facial features and their orthodontic implications. Am J
Orthod and Oral Surg 1939;25:927-951.
41. Howells DJ, Shaw WC. The validity and reliability of ratings of dental and facial
attractiveness for epidemiologic use. Am J Orthod 1985;88:402-408.
53
42. Iliffe AH. A study of preferences in feminine beauty. Br J Psychol 1960;51:267-
273.
43. Keating CF. Gender and the physiognomy of dominance and attractiveness. Soc
Psych Quar 1985;48:61-70.
44. Kerr WJ, O’Donnell JM. Panel perception of facial attractivess. Br J Orthod
1990;17.
45. Kiekens RM, Maltha JC, van't Hof MA, Kuijpers-Jagtman AM. A measuring
system for facial aesthetics in Caucasian adolescants: reproducibility and validity.
Eur J Orthod 2005;27:579-584.
46. Kiekens RM, Maltha JC, van't Hof MA, Kuijpers-Jagtman AM. Objective
measures as indicators for facial esthetics in white adolescents. Angle Orthod
2006;76:551-556.
47. Kilpelainen PV, Phillips C, Tulloch JFC. Anterior tooth position and motivation
for early treatment. Angle Orthod 1993;63:171-174.
48. Kiyak HA, Zeitler DL. Self-assessment of profile and body image among
orthognathic surgery patients before and two years after surgery. J Oral Maxillofac
Surg 1988;46:365-371.
49. Knight H, Keith O. Ranking faical attractiveness. Eur J Orthod 2005;27:340-348.
50. Langlois JH, Kalakanis LE, Rubenstein AJ, Larson AD, Hallam MJ, Smoot MT.
Maxims or myths of beauty? A meta-analytic and theoretical overview. Psychol Bull
2000;126:390-423.
51. Langlois JH, Roggman LA. Attractive faces are only average. Psychol Sci
1990;1:115-121.
52. Lew KK, Soh G, Loh E. Ranking of facial profiles among Asians. J Esthet Dent
1992;4:128-130.
53. Lines PA, Lines RR, Lines CA. Profilemetrics and facial esthetics. Am J Orthod
1978;73:648-657.
54. Ludlow JB, Davies-Ludlow LE, Brooks SL, Howerton WB. Dosimetry of 3
CBCT devices for oral macillofacial radiology: CB Murcuray, NewTom 3G and i-
CAT. Dentomaxillofac Radiol 2006;35:219-226.
55. Lundstrom A, Popovich F, Woodside DG. Panel assessment of the facial frontal
view as related to mandibular growth direction. Eur J Orthod 1989;11:290-297.
54
56. Lundstrom A, Woodside DG, Popovich F. Panel assessment of facial profile
related to mandibular growth direction. Eur J Orthod 1987;9:271-278.
57. Mackley RJ. An evaluation of smiles before and after orthodontic treatment.
Angle Orthod 1993;63:183-189.
58. Mah JK, Danforth RA, Bumann A, Hatcher D. Radiation absorbed in
maxillofacial imaging with a new dental computed tomography device. Oral Surg
Oral Med Oral Pathol Oral Radiol Endod 2003;96:508-513.
59. Mah JK, Hatcher D. Current status and future needs in craniofacial imaging.
Orthod Craniofacial Res 2003;6:10-16.
60. Maple JR, Vig KW, Beck FM, Larsen PE, Shanker S. A comparison of providers'
and consumers' perceptions of facial-profile attractiveness. Am J Orthod Dentofac
Orthop 2005;128:690-696.
61. Martin JG. Racial ethnocentrism and judgment of beauty. J Soc Psychol
1964;63:59-63.
62. Matoula S, Pancherz H. Skeletofacial morphology of attractive and nonattractive
faces. Angle Orthod 2006;76:204-210.
63. Maxwell R, Kiyak HA. Dentofacial appearance: a comparison of patient self
assessment techniques. Int J Adult Orthod Orthognath Surg 1991;6:123-131.
64. Meehan M, Teschner M, Girod S. Three-dimensional simulation and prediction
of craniofacial surgery. Orthod Craniofacial Res 2003;6:102-107.
65. Mew J. Suggestions for forecasting and monitoring facial growth. Am J Orthod
Dentofac Orthop 1993;104:105-120.
66. Michiels G, Sather AH. Validity and realiability of facial profile evaluation in
vertical and horizontal dimensions from lateral cephalograms and lateral
photographs. Int J Adult Orthod Orthognath Surg 1994;9:43-54.
67. Miller AG. Role of physical attractiveness in impression formation. Psychonom
Sci 1970;19:241-243.
68. Moss JP, Coombes AM, Linney AD, Campos J. Methods of three dimensional
analysis of patients with asymmetry of the face. Proc Finn Dent Soc 1991;87:139-
149.
69. Moss JP, Linney AD, Lowey MN. The use of three-dimensional techniques in
facial esthetics. Semin Orthod 1995;1:94-104.
55
70. Naini FB, Moss JP. Three-dimensional assessment of the relative contributions
of genetics and enviornment to various facial parameters with the twin method. Am J
Orthod Dentofac Orthop 2004;126:655-665.
71. Naini FB, Moss JP, Gill DS. The enigma of facial beauty: esthetics, proportions,
deformity, and controversy. Am J Orthod Dentofac Orthop 2006;130:277-282.
72. Peck H, Peck S. A concept of facial esthetics. Angle Orthod 1970;40:284-318.
73. Peck S, Peck L, Kataja M. Skeletal asymmetry in estetically pleasing faces.
Angle Orthod 1991;61:43-48.
74. Penton-Voak IS, Perrett DI, Castles DL, Kobayashi T, Burt DM, Murray LK,
Minamisawa R. Menstrual cycle alters face preference. Nature 1999;399:741-742.
75. Perrett DI, May KA, Yoshikawa S. Facial shpare and judgements of female
attractiveness. Nature 1994;368:239-242.
76. Perseo G. The "beauty" of homo sapiens: standard canons, ethnical, geometrical
and morphological facial biotypes. An explained collection for frontal north-
europide comtemporary beauty facial canons. part 1. Virtual J Orthod 2002;30:150-
162.
77. Phillips C, Griffin T, Bennett E. Perception of facial attractiveness by patients,
peers, and professionals. Int J Orthod Orthognath Surg 1995;10:127-135.
78. Phillips C, Trentini CJ, Douvartzidis N. The effect of treatment on facial
attractiveness. J Oral Maxillofac Surg 1992;50:590-594.
79. Phillips C, Tulloch C, Dann IV C. Rating of facial attractiveness. Comm Dent
Oral Epidemiol 1992;20:214-220.
80. Pogrel MA. What are normal esthetic values? J Oral Maxillofac Surg
1991;49:963-969.
81. Powell N, Humphreys B. Proportions of the aesthetic face. New York: Thieme-
Stratton, Inc; 1984.
82. Powell SJ, Rayson RK. The profile in facial aesthetics. Br J Orthod 1974;3:207-
215.
83. Prahl-Andersen B, Boersma H, van der Linden FP, Moore AW. Perceptions of
dentofacial morphology by laypersons, general dentists, and orthodontists. JADA
1979;98:209-212.
56
84. Proffit WR, Phillips C, Douvartzidis N. A comparison of outcomes of
orthodontic and surgical-orthodontic treatment of Class II malocclusion in adults.
Am J Orthod Dentofac Orthop 1992;101:556-565.
85. Proffit WR, White RP, Sarver DM. Contemporary treatment of dentofacial
deformity. St Louis: Mosby; 2003.
86. Ras F, Habets LL, van Ginkel FC, Prahl-Andersen B. Method for quantifying
facial asymmetry in three dimensions using stereophotogrammetry. Angle Orthod
1995;65:233-239.
87. Ras F, Habets LL, van Ginkel FC, Prahl-Andersen B. Three-dimensional
evaluation of facial asymmetry in cleft lip and palate. Cleft Palate Craniofac J
1994;31:116-121.
88. Ricketts RM. Divine proportion in facial esthetics. Clin Plast Surg 1982;9:401-
422.
89. Riedel RA. An analysis of dentofacial relationships. Am J Orthod 1957;43:103.
90. Riedel RA. Esthetics and its relation to orthodontic therapy. Angle Orthod
1950;20:168-178.
91. Rigsbee III OH, Sperry TP, BeGole EA. The influence of facial animation on
smile characteristics. Int J Adult Orthod Orthognath Surg 1988;3:233-239.
92. Rivera SM, Hatch JP, Dolce C, Bays RA, Van Sickels JE, Rugh JD. Patients'
own reasons and patient-percieved recommendations for orthognatiic surgery. Am J
Orthod Dentofac Orthop 2000;118:134-141.
93. Romani KL, Agahi F, Nanda R, Zernik JH. Evaluation of horizontal and vertical
differences in facial profiles by orthodontists and lay people. Angle Orthod
1993;63:175-182.
94. Scott CR, Goonewardene MS, Murray K. Influence of lips on the perception of
malocclusion. Am J Orthod Dentofac Orthop 2006;130:152-162.
95. Shaw WC. The influence of children’s dentofacial appearance on their social
attractiveness as judged by peers and lay adults. Am J Orthod 1981;79:399-415.
96. Shaw WC, Rees G, Dawe M, Charles C. The influence of dentofacial appearance
on the social attractiveness of young adults. Am J Orthod 1985;87:21-26.
97. Shelly AD, Southard TE, Southard KA, Casko JS, Jakobsen JR, Fridrich KL,
Mergen JL. Evaluation of profile esthetic change with manidibular advancement
surgery. Am J Orthod Dentofac Orthop 2000;117:630-637.
57
98. Soh J, Chew MT, Wong HB. A comparative assessment of the perception of
Chinese facial profile esthetics. Am J Orthod Dentofac Orthop 2005;127:692-699.
99. Steiner CC. Cephalometrics for you and me. Am J Orthod 1953;39:720-755.
100. Tatarunaite E, Playle R, Hood K, Shaw W, Richmond S. Facial attractiveness: a
longitudinal study. Am J Orthod Dentofac Orthop 2005;127:676-682.
101. Tedesco LA, Albino JE, Cunat JJ, Green LJ, Lewis EA, Slakter MJ. A dental-
facial attractiveness scale. Part I. Reliability and validity. Am J Orthod 1983;83:38-
43.
102. Tedesco LA, Albino JE, Cunat JJ, Slakter MJ, Waltz KJ. A dental-facial
attractiveness scale. Part II. Consistency of perception. Am J Orthod 1983;83:44-46.
103. Todd SA, Hammond P, Hutton T, Cochrane S, Cunningham S. Perceptions of
facial aesthetics in two and three dimensions. Eur J Orthod 2005;27:363-369.
104. Tulloch C, Phillips C, Dann IV C. Cephalometric measures as indicators of
facial attractiveness. Int J Orthod Orthognath Surg 1993;8:171-179.
105. Tuominen ML, Tuominen RJ, Nystrom ME. Subjective orthodontic treatment
need and perceived dental appearance among young Finnish adults with and without
previous orthodontic treatment. Comm Dent Health 1994;11:29-33.
106. Udry J. Structural correlates of feminine beauty preference in Britian and the
U.S.: a comparison. Sociol Soc Res 1965;49:330-342.
107. Vargo JK, Gladwin M, Ngan P. Association between ratings of facial
attractiveness and patients’ motivation for orthognathic surgery. Orthod Craniofacial
Res 2003;6:63-71.
108. Wilmot JJ, Barber HD, Chou DG, Vig KW. Associations between severity of
dentofacial deformity and motivation for orthodontic-orthognathic surgery treatment.
Angle Orthod 1993;63:283-288.
109. Wood TJ. Aesthetics and orthodontics. Dent Practit 1969;19:247-260.
110. Zaidel DW, Cohen JA. The face, beauty, and symmetry: perceiving asymmetry
in beautiful faces. Int J Neuroscience 2005;115:1165-1173.
58
Appendix 1
To: All USC, School of Dentistry students and Employees
Date: January, 2007
RE: Request to participate in a study rating FACIAL ATTRACTIVENESS.
STUDY: Comparative facial attractiveness in two and three dimensions using
archived images
PRINCIPAL INVESTIGATOR: Dr. James Mah
CO-INVESTIGATOR: Dr. Jordan Millar and Dr. Glenn Sameshima
STUDY CO-ORDINATOR: Dr. Jordan Millar
We are conducting research on two-dimensional (2D) and three-dimensional (3D)
facial attractiveness and require orthodontists, orthodontic residents, dental students
and general public to participate in rating the images. We will require approximately
30-45 minutes to complete the study. The study coordinator will contact you by
email to confirm the date, time, and location of the study.
For more information or if you would like to participate, please contact the study
coordinator, Dr. Jordan Millar, at jmillar@usc.edu or in room 318.
Thank you,
Jordan Millar, DMD
Study Coordinator
59
Appendix 2
STUDY: Comparative facial attractiveness in two and three dimensions using
archived images
PRINCIPAL INVESTIGATOR: Dr. James Mah
CO-INVESTIGATOR: Dr. Jordan Millar and Dr. Glenn Sameshima
STUDY CO-ORDINATOR: Dr. Jordan Millar
Rater Agreement
January, 2007
This study involves looking at faces in two-dimensions (2-D) (photographs) and three-
dimensions (3-D) (video clip) and consists of two parts. It will take approximately 30
minutes.
You will not be given any specific information about the faces you will see. Your only task
is to assess the facial attractiveness of each. There is no right or wrong answer. We ask that
you make your decisions independently. You will not be allowed to return to a previous
image once you have begun.
A series of two introductory images (one 2-D image and one 3-D video) will be shown to
accustom you to the procedure. Each image will be shown for 12 to 25 seconds with a period
of 4 seconds where no image is shown to allow time for scores to be recorded.
PART 1:
You will be scoring 40 images on a scale of 0 (least attractive) to 100 (most attractive). Each
image will be shown for 12 to 25 seconds with a period of 4 seconds where no image is
shown to allow time for scores to be recorded.
PART 2:
Five to six images will be shown simultaneously and you will rank them from left (most
attractive) to right (least attractive). Two-D images will be shown for 45 seconds and 3-D
images will be shown for 60 seconds. Males and females will be viewed and ranked
separately. You cannot assign the same ranking to more than one image.
I, _______________________________(print) agree to participate in the study entitled
“Comparative facial attractiveness in two and three dimensions using archived images”. The
study will be conducted by Dr. James Mah, Dr. Jordan Millar, and Dr. Glenn Sameshima.
I agree to protect the confidentiality of the patients in this study by not discussing the study.
______________________________ ______________________________
Signature of Rater: Signature of Study Coordinator:
______________ _______________ ______________________________
Date: Age: Date:
60
______________ ___M_____F____
Ethnicity: Gender (circle)
Please circle one of the following groups: DENTAL STUDENT GENERAL
PUBLIC
1
st
yr 2
nd
yr 3
rd
yr 4
th
yr
ORTHODONTIST ORTHODONTIC
RESIDENT
Abstract (if available)
Abstract
Introduction: The purpose of this study was to compare facial esthetic preferences in two-dimensions (2-D) and three-dimensions (3-D) and to determine the preferred facial relationship. Methods: Orthodontists (n=33), dental students (n=26), and laypersons (n=27) assessed 2-D images (frontal, profile) and 3-D surface images of 20 subject within three facial profile: Class I (n=9), Class II (n=7), and Class III (n=4) on a visual analog scale. A subset from each profile type was ranked in order of preference for 2-D and 3-D images separately. The data was analyzed using univariate and multivariate statistics. Results: Three-D images are viewed significantly more attractive than 2-D. Professionals and laypeople preferred the Class I facial profile in 2-D and 3-D. However, orthodontists rated significantly higher than laypeople and dental students. Conclusions: Although the clinical significance of these differences is unknown, 3-D surface imaging is a tool that will bring orthodontic records into the next generation.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Comparative study of Caucasian and Hispanic mandibular clinical arch forms
PDF
A comparative study of caucasian and African American mandibular clinical arch forms
PDF
Comparison between tooth mesiodistal angulation measurements from constructed panoramic images and three-dimensional volumetric images
PDF
Proper mesio-distal angulation and bucco-lingual inclination of the whole tooth in three-dimensional space -- a standard for orthodontic patients
PDF
A three-dimensional computed tomography comparison of the maxillary palatal vault between patients with rapid palatal expansion and orthodontically treated controls
PDF
Slow maxillary expansion for the treatment of unilateral crossbite in pre-adolescents: a long-term retrospective study of the changes in arch dimension
PDF
The correlation between dental caries in primary and permanent teeth in children at two different elementary schools
PDF
The relationship of the forehead to the maxillary central incisor in adult white females: an evaluation of Andrews Element II analysis
PDF
Relationship between dental and alveolar bony arch form and whole tooth mesiodistal angulation and faciolingual inclination in three-dimensional space
PDF
Comparison of premolar extraction rates between one-phase and two-phase class II malocclusion
PDF
A cone beam CT evaluation of the maxillary dento skeletal complex after rapid palatal expansion
PDF
Relationship between key cephalometric parameters and tooth tip and torque
PDF
Orthodontic confidence of senior dental students: a study of 2 US dental schools
PDF
Comparison of cortical bone thickness between second premolars and first molars in the maxilla and mandible in four ethnic groups
PDF
The mesiodistal angulation and faciolingual inclination of each whole tooth in three dimensional space post non-extraction orthodontic treatment
PDF
Effect of cyclooxygenase inhibitors on rat root resorption and tooth movement
PDF
Prevalence of arch form types and tooth size discrepancy in a Vietnamese population
PDF
Utilizing voxel based superimposition to asses orthognathic surgical treatment
PDF
The effect of cone beam computed tomography (CBCT) imaging on orthodontic diagnosis and treatment planning
PDF
Cephalometric norms for the Vietnamese: a clinical appraisal
Asset Metadata
Creator
Millar, Jordan
(author)
Core Title
Comparative facial attractiveness in two and three dimensions
School
School of Dentistry
Degree
Master of Science
Degree Program
Craniofacial Biology
Publication Date
03/28/2007
Defense Date
03/12/2007
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
facial attractiveness,OAI-PMH Harvest
Language
English
Advisor
Sameshima, Glenn T. (
committee chair
), Enciso, Reyes (
committee member
), Mah, James (
committee member
)
Creator Email
jmillar@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m336
Unique identifier
UC1106757
Identifier
etd-Millar-20070328 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-321839 (legacy record id),usctheses-m336 (legacy record id)
Legacy Identifier
etd-Millar-20070328.pdf
Dmrecord
321839
Document Type
Thesis
Rights
Millar, Jordan
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
facial attractiveness