NTC Project: S01-AC27 (formerly 101-27)

Body Scan Analysis for Fit Models Based on Body

Shape and Posture Analysis
Lenda Jo Connell, Principal Investigator, Pam Ulrich, Auburn University, Alistair
Knox and Graham Hutton, Nottingham Trent, Dave Woronka,[TC]² , Susan
Ashdown, Cornell

Objective:
Recently a number of retailers have experienced a down turn in growth and lower
profit margins (Reuters, 8/11/2000). On a whole, the mid-priced apparel market
appears to be saturated with “safe” merchandise lacking variety, styling and fit.
Studies by Kurt Salmon Associates and others (1998, 1999) have consistently
shown that one problem is that more than half of the female population cannot
find apparel in the marketplace to fit. One problem with the fit of women’s apparel
is that it is based on outdated data. The purpose of this study is 1) to analyze
body scan data from women 19-55 relative to body shape, posture, and weight to
develop body analysis standards, 2) to produce an expert system which can be
used to analyze body scans as a basis for realigning sizing and pattern
development for specific target markets, and 3) to develop virtual fit models and
slopers based on body shapes and posture occurring in the population to
enhance the fit of mass produced women’s apparel.

Abstract:
Body shape analysis is the theoretical underpinning for the development of
apparel sizing. Currently based on standardized sizing, this shape analysis
builds a garment structure that must fit on many body shapes within a size range
as standardized sizes are not currently made to accommodate body shapes
differing from the standard. Most sizing systems used in the U.S., are based on
an hourglass figure shape that has a proportionately smaller waist (8-10”) when
compared to the hip/bust measurement. Data available through body scanning
now allows researchers to investigate body shape in depth. This study is
focused on body shape as presented in the U. S. population and relevant for
understanding sizing and fit for women. A methodology identifying the shape of
eight body components important in apparel sizing have been developed. The
components include body build, body shape, hip shape, shoulder slope, front
torso shape, buttocks shape, back curvature, and posture. The methodology is
being tested with shape experts. Data from field tests with shape analysis
experts will be used to develop software to classify body shape.

Background:

Studies by Kurt Salmon Associates and others (1998, 1999) have consistently
shown that one problem with slow apparel sales is that more than half of the
female population cannot find apparel in the marketplace to fit. The fit of

1
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
2
women’s apparel is based on outdated data. The purpose of this study is to
better understand body shape as an element in defining better basic fit models
for pattern development.
Douty (1963) first contributed to the basic research on body shape relative
to fit using a research technique called somatography. Using back-lit
photography, she projected bodies of subjects on a grid to study body shape and
posture. She developed rating scales for body shape and posture. However,
Douty’s figures were based on an hourglass shape developed incrementally on
both the top and bottom of the figures.
Other researchers have proposed portions of body scales but no analytical
scale exists in the public domain to allow product developers to compare their
target market to a standard shape as a basis for understanding pattern
production and grading. Today body scan technology allows researchers to look
at the body in very different ways relative to pattern development for apparel
production. Some body scanning centers have developed shape analysis tools
for use in pattern development. Nottingham Trent uses both a 2-D and a 3-D
program to identify body shape. [TC]² has begun to work on a program for shape
analysis. However, most of this work has been done under proprietary
conditions and is not available for academic scrutiny. Some of the work may
have been done with a technical approach to modeling. The International
Standard Organization (ISO) revised their sizing system in 1991 to reflect
differences in body shape. They use drop measurements, the difference
between hip circumference and bust circumference), to sort women into body
shape categories classified as A, M and H body types. Though this work is a
beginning of recognition that a figure outside of the hourglass exists; it is not a
comprehensive assessment of body shape. And much of this work has been
completed by computer scientists and modeling programs.
Earlier survey research within the NTC (Anderson, Brannon, Ulrich,
Presley, 1998) based on self-reports of female consumers indicated that body
shape varied with age and ethnicity. While 53% of college-aged females
reported having an hourglass body shape, only 33% of female in a national study
that included a broader age range reported having an hourglass shape. This is
important because apparel patterns made for an hourglass figure are not shaped
to fit bodies may have wider waists or hips.
Body analysis for pattern development extends beyond just body shape.
Fit experts understand the relationship of the flat pattern that is used to fit apparel
to the curve of the body that the pattern must fit. Douty (1963) early on identified
posture as an important factor in fit. Posture that is erect tends to flatten body
curves. Posture that is compensating produces curves particularly in the
buttocks and shoulders that must be accounted for to make garments fit female
bodies. In addition to body shape and posture, body build, front torso shape,
shoulder shape, hip shape, back curvature, and buttocks shape are all defining
points for pattern development for the female figure.

2
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
3
Project Approach:

NTC researchers have used body scan data collected from females ages
19-55 to develop a tool to aid in expert analysis of body shapes. Researchers,
all experts in fit, used the Douty (1963) scales as a launching point for
understanding body shape. They identified critical body areas for adjusting
patterns for fit of apparel. Body shapes over the group of body scans in the
study were analyzed for variances from a core figure representing an industry
defined average figure type. Industry uses the hourglass figure as a standard fit
model from which patterns for target markets in women’s wear are developed.
This project has developed the NTC Women’s Body Shape Analysis Tool
(WBSAT) which uses the hourglass standard as a launching point for
understanding body shape but identifies the three other predominant body
shapes for women including the rectangle, pear and inverted triangle shape.
Figures 1,2, and 3 show various figure components in the NTC Guide for
Women’s Body Shape Analysis Tool. The guide consists of line drawings of
eight figure scales which can be used to assess body build, body shape, posture,
hip shape, front torso shape, buttocks shape, back curvature, and shoulder
slope. These scales were developed from figures existing within body scan data
and agreed on by experts to be characteristic of the figure trait being evaluated.
Body scan data has been collected from over 400 females in the U.S. and
the U.K. To establish validity and reliability for the WBSAT, the scales will be
tested with experts in fit from academia and industry. Five experts qualified by
experience in teaching and practicing pattern development and fit will use the
tool to assess 100 body scans on the eight scales within the tool. Inter-rater
reliability will be calculated. Experts will be interviewed for feedback on the ease
of use and accuracy of the scales in classifying body shapes within the sample.
Once the 100 body scans have been rated using the WBSAT, inter-rater
reliability will be calculated. After establishing reliability, researchers will work
with [TC]² to train software for body shape analysis using the body scans rated
by the team of fit experts. [TC]² has developed a shape analysis tool but it needs
to be populated with valid data for defining shape. This approach allows the
body shape analysis software to be defined by both technical and discipline
specific experts.
After software is developed to analyze the shape component of body
scans, researchers at Auburn will use it to evaluate the remaining existing body
scans seeking to understand critical relationships within body shape that may
have an impact on fit. Relationships may include some of the following:
Are there hip shapes that are related to body shape?
Is posture related to buttocks shape or shoulder curvature?
Is body shape related to body build?
Is torso shape related to posture or body build?
Is torso shape related to hip shape?
Is torso shape related to body shape?
Is shoulder shape related to back curvature?

3
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
4
Figure 1: Body Shape Analysis

1 2 3 4
Hourglass Rectangular Pear Inverted Triangle

Shoulder and hip width Approximately equal Much wider at hip than Much wider at shoulder
equal or nearly equal with visual balance between shoulder with waist than hip with waist
very small waist shoulder and hip with indented. Overall indented. Overall
indentation. Overall waist slightly indented. appearance of appearance of extremely
appearance of extremely Overall appearance of extremely heavy hip broad shoulder relative
area relative to waist
small waist relative to
shoulder and hip.
average figure with
little waist definition. and shoulder. to waist and hip .

4
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
5
Figure 2: Torso Analysis

Core Figure Moderate B Heavy B

CF B-1 B-1

Figure has flat Indented waist with Slightly indented waist

indented waist and moderate fullness above with extreme fullness
slightly extended and below above and below
abdomen.

5
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
6

Figure 3: Body Build: Overall appearance of body size, ranging from thin to full-figured

6
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
7

4
1 2 3 Heavy
Thin Average Full
Substantial width or
Torso, arms & legs Torso, arms & legs have Torso, arms & legs fullness through
appear thin, perhaps moderate, soft or appear fuller or torso, upper legs &
bony or caved in; muscular padding over wider; upper arm & arms; waist is less
waist may be defined skeleton, which may not torso (or parts of likely to be clearly
or straight. be evenly distributed torso) may appear defined; fullness may
throughout body; waist particularly fuller. not be evenly
may be defined or distributed.
straight.

[TC]² is currently involved with the Size USA project. As a member of the
consortium, fit researchers with this project will have access to a broader range
of data.
Nottingham Trent will be involved in development and testing of the shape
analysis tool. They have an existing shape analysis program that can be used to
test the WBSAT tool. Researchers will be able to compare any variances in the
shape analysis programs and actual shape analysis data from the U.K. with
shape data from U.S. populations.
In addition the body shape data, researchers have also collected data on
body cathexis (how women feel about their bodies), how women prefer
categories of apparel jeans, pants, jackets, dresses) to fit their bodies, what
benefits women seek from clothing (figure flaw compensation, sex appeal,
fashion innovation, and fashion Image), where women prefer to shop, areas
where fit problems exist, and basic demographic information. Once distinct body
shapes are identified, these data will be analyzed relative to body shape.
In addition to developing the shape analysis tools, this project will be able
to build virtual fit models based on actual shape data available from body scan
data. Fit models are prototype bodies on which apparel sizing designations are
based. Each manufacturer defines fit for their target market. Their definition may
differ from other manufacturers. Fellingham (1991) reported on a size 8 fit model
who was 5’8”, weighed 115 pounds and reported body measurements of 36-24-
34. When this fit model tried on six different size 8 sheaths, the result was a
dramatic difference in fit for each and indicated the lack of consistency among
manufacturers in defining even one size. Analysis of the fit models
measurements in even a cursory manner would indicate problems in fitting a
mass market with sizing based on a fit model who is taller, probably weighs less
than, and may have a significantly smaller hip measurement than an average
size 8 in the general population. Data collected from a national study of over
1000 women indicates that the average woman in the U.S. today is nearly 5’5”
tall, weighs 159 pounds and measures 37.3, 31.7 and 37.6 at the bust, waist and
hips. A virtual model of this figure based on these measurements can be seen in
Figure 4.

7
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
8
Figure 4: Average female figure in 19-55 target market

National Textile Center

Average Female Figure

Once fit models are defined, pattern development will begin based on
emerging body shapes and classifications. Virtual models can be used to display
and analyze fit.

Contributing Graduate Students: Ma Li, Gina Pisut, Mei-Lin Fu, Juan Juan
Wu, Lara Belliston, Marina Alexander

Theses In Process:
Li, M. (2002, Fall). An exploratory study of body shape and body build to
understand women’s fit preferences and problems. Auburn University.

Dissertations in Process:

Alexander, M. Understanding body shape patterns from body scan data. Auburn
University.

Fu, M. Body shape analysis for Taiwanese female figures . Auburn University.

Invited Publications:

Kaplan, A. (2002, February 13), Try this on for sighs. St. Paul Pioneer Press, E,
10.

8
National Textile Center Annual Report: November 2002
 NTC Project: S01-AC27 (formerly 101-27)
9
McInerny, V. (2001, January). Having a fit. The Oregonian, L, 1.

Presentations:

Connell, L.J., Brannon, E.L., and Ulrich P.V. (2002). A consumer driven model
for mass customization in the apparel market. Journal of Fashion Marketing and
Management, 6(3), 240-256.

Pisut, G. and Connell**, L. J. ( 2001, November). Fit preferences of female

consumers relative to body shape, body cathexis, clothing benefits sought, and fit
problems: a national study. To be presented at the International Textile and
Apparel Association Annual Conference and published in the Proceedings of the
International Textile and Apparel Association . Kansas City, MO

Connell**, L. J., Brannon, E., Ulrich, P., and Presley, A. B. (2001, June)
Exploring female consumers' fit preferences. Presented at the Korean Society of
Clothing and Textiles /International Textile and Apparel Association Joint World
Conference and published in the Proceedings of the International Textile and
Apparel Association. Seoul, Korea.

Copyrights: NTC Women’s Body Shape Analysis Tool; NTC Guide for
Women’s Body Analysis

Industry Contacts: JC Penney, Lands’ End, Dillard’s, SAK’s, PAD, Judy Lynn

Other University Contacts: Cornell, Oregon State, Radford, Nottingham Trent

Web Site: www.humsci.auburn.edu/ca/faculty/faculty.htm

9
National Textile Center Annual Report: November 2002