Original article

Textile Research Journal

2018, Vol. 88(15) 1721–1734

Design and evaluation of personalized ! The Author(s) 2017

Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
garment block for atypical morphology DOI: 10.1177/0040517517708537
journals.sagepub.com/home/trj
using the knowledge-supported virtual
simulation method

Yan Hong1,2,3, Pascal Bruniaux2, Xianyi Zeng2,

Antonella Curteza3 and Kaixuan Liu4

Abstract
This research puts forward a novel knowledge-supported design process for obtaining personalized ready-to-wear
garment products aimed at consumers with atypical morphology by using a virtual a three-dimensional-to-two-
dimensional (3D-to-2D) design method. The proposed design process starts with designing a personalized garment
block, which is then extended into the desired ready-to-wear garment style. The garment block is obtained by using a
virtual 3D-to-2D design method. The extension and sizing of the garment block pattern into desired ready-to-wear
garment patterns is performed in a 2D environment using classic methods. The proposed design solution begins with a
personalized garment block, thus avoiding the complicated operations of simulating a 3D garment directly in the virtual
environment. The proposed design process can be fully digitalized, which ensures the involvement of the consumer
throughout the design. By repeatedly running the sequence of Design – Display – Evaluation – Adjustment, which is
technically accomplished through the virtual 3D-to-2D design method and 3D virtual try-on, the designers’ patterns
and principles of elaborating personalized garment products can be validated within a very short time. As a knowledge-
supported design process, designers’ design ideas and principles of personalized design solution can be fully extracted to
enhance satisfaction of the final product. A set of experiments and case studies validated the approach and application of
the proposed knowledge-supported virtual design process, and demonstrated its efficiency.

Keywords
virtual design process, three-dimensional garment design, knowledge-supported design, subjective visual evaluation, fuzzy
logic, atypical morphology

In response to the quickly increasing consumer market
for individual needs, customized design and production
have been recognized as effective means to meet cus-
tomers’ needs, improving individual satisfaction.1
Fashion, as an important category of consumer prod-
ucts, has been considered as one of the industrial prod-
uct types that should find a solution for customized
design and production.2 The rising requirements of cus-
tomized/personalized design in the fashion industry
strengthen the importance of the personalization ser-
vice, especially for consumers with particular require-
ments, such as atypical morphology.3
There are many individuals within the categories of
atypical morphology, such as the elderly, impaired,
and/or disabled, who have exceptional problems with
finding suitable clothes. This group requires persona-
lized products concerning sizes, styles, needs, or expres-
sions.4 However, these individuals have unique figures,
which do not fit into the standard sizes that are avail-
able in stores. The current design solution for this
1
College of Textile and Clothing Engineering, Soochow University, China
2
Ecole Nationale Superieure des Arts et Industries Textiles, France
3
Gheorghe Asachi Technical University of Iasi, Romania
4
Dong Hua University, China
Corresponding author:
Yan Hong, College of Textile and Clothing Engineering, Soochow
University, China, No.1 Shizi Street, Suzhou, Suzhou 215021, China.
Email: yannichonghk@gmail.com
1722
group is the adaption of existing garment patterns.
However, this solution is not only time consuming
but also leads to a high cost.5 Also, it is not possible
to offer customers with atypical morphology a garment
to order, according to their desires and body figure.6
There is a gap between the stores’ supply and the
demands from customers. The more the figure diverges
from the standard figure, the more difficult it is to find
suitable clothes in retail stores.7
Recently, with the development of the three-
dimensional (3D) laser body scan, surface recon-
struction, and virtual reality technology, the virtual
three-dimensional-to-two-dimensional (3D-to-2D) design
method has emerged as a new solution to personaliza-
tion service in the fashion industry.8 Using the virtual
3D-to-2D design method, a personalized garment is
designed directly on a 3D scanned body and developing
corresponding garment patterns emerges as a research
interest for the fashion industry.5 Due to the develop-
ment of the virtual 3D-to-2D design method, the man-
ufacturing process of product design and development
(PD&D) is changing in the fashion and textile industry,
particularly in the domain of personalized garment
product design.9 As a novel design solution, the virtual
3D-to-2D design method is expected to be able to
effectively satisfy the personalized garment design
requirements.
Compared to some traditional design methods, such
as traditional 3D-to-2D and 2D-to-3D methods, the
virtual 3D-to-2D design method is able to integrate
the concept of collaborative design in the PD&D
process.10,11 By running the sequence of Design –
Display – Evaluation – Adjustment repeatedly, which
is technically accomplished by applying the virtual
3D-to-2D design method and 3D virtual try-on, the
whole process can be presented to both consumers
and designers in real time.12 For example, consumers
can be involved in the whole design process by using
a remote monitor. Also, evaluation based on the dis-
played design result can be performed by both
designers and consumers. Evaluation results constitute
the input of the adjustment section. Designers’ design
ideas and principles of the personalized design solution
can be validated within a very short time through this
design sequence.
The novel PD&D process using the concept of the
virtual 3D-to-2D method not only changes the working
process, but also leads to a totally new business model
for the increasing market of atypical morphology. Due
to the complexity of designing for atypical morphology,
a consumer’s satisfaction can be greatly increased
by his/her involvement in the collaborative design pro-
cess. By using the evaluation section, the designer’s
knowledge can be extracted to support the design
process and enhance the consumer’s satisfaction.13
Textile Research Journal 88(15)
Fashion firms are able to win a greater share of the
market through the development of a more advanced
business model based on the application of the virtual
3D-to-2D design process.
In this context, the virtual 3D-to-2D design method
seems to be the ideal solution for atypical morpholo-
gies. Traditionally, most of the research about 3D-to-
2D virtual design has been focused on the development
of 3D garments directly from the scanned body; how-
ever, due to the diversity of ready-to-wear garment
styles, 3D garment simulation requires not only high-
level technical support for the software, but also an
experienced designer to perform the design process.14
It is difficult to extract systematic design rules for con-
ceiving ready-to-wear products of all kind of styles
using the virtual 3D-to-2D design method, which is
contrary to the mainstream trend of automatic design
in the computer-aided design (CAD) area.15 To avoid
these problems, some researchers have tried to develop
personalized garment block patterns using a virtual 3D-
to-2D design method, which can be further extended
into accurate 2D garment patterns for ready-to-wear
products of desired styles. Time, labor, and technical
expense can be greatly reduced if this method is used.
For example, recently, Hong, Zeng and Bruniaux5
developed a garment block design method by defining
a set of landmarks along with corresponding wearing
ease controlling lines on the scanned body surface.
Using surface reconstruction technology, a persona-
lized garment block pattern can be obtained.16
However, the precision of the design effect is largely
dependent on the quantity of the defined landmarks.
The wearing ease of each landmark should be given
manually, which is also rather complex. Thus, a more
efficient 3D-to-2D design method is required for gar-
ment block generation.
In this research, a virtual 3D-to-2D-based design
process is proposed for designing personalized ready-
to-wear products aiming at atypical morphology. The
proposed design process begins with designing a perso-
nalized garment block, which can be further extended
into the desired style for the related ready-to-wear
product. Obtaining the personalized garment block is
realized by using the proposed virtual 3D-to-2D design
method, which drapes the shape of the garment block
on the scanned human body directly, in a 3D virtual
environment. The extension and sizing of the garment
block pattern into desired ready-to-wear garment pat-
terns is carried out in a 2D environment using classic
methods. The garment block, regarded as a primary
design of the ready-to-wear product, has all the essen-
tial elements for a garment, such as the construction,
opening, darts, and controlled ease, without extra style
information.17 Comparing the style of ready-to-wear
products, which is rather subjective, an agreement for
Hong et al.
the general shape of a garment block can be easily
reached by designers without too many conflicts.18
In this study, the research domain is mainly focused
on the introduction of the development of a sleeveless
garment block, which can be fast applied to atypical
morphology. The design and construction of the sleeves
are not involved in the proposed design method,
because by detecting the circumference of different arm-
hole curves on the obtained garment block body,
sleeves can be easily obtained using the 2D classic pat-
tern design method. This will certainly reduce the com-
plexity of the proposed design method. An evaluation is
performed by both designers and consumers, based on
the design result obtained from the proposed method.
During the procedure of subjective visual evaluation,
the design knowledge and experience of designers will
be extracted to provide solutions in the case of an
unsatisfactory design effect. Using this method, the
concept of knowledge-support design is applied so as
to enhance consumers’ satisfaction.
The rest of the study is structured as follows. The
general working process of the proposed garment block
development using the virtual 3D-to-2D design method
along with the related software is introduced in
the Methodology section. In the Experiment details sec-
tion, four experiments are introduced to thoroughly
explain an application of the proposed garment block
design process. Twenty design cases are presented.
Furthermore, validation experiments are also depicted
in this section with the purpose of investigating if the
proposed design process could be efficiently accom-
plished for atypical morphology while ensuring con-
sumer satisfaction. Also, it is validated that the design
knowledge of experts can be extracted to support the
design process. The third section is concerned with the
differences between the initial and the modified garment
block pattern using a set of key fit measurements
(KFMs). The final section concludes this study.
Methodology
Virtual design platform
In this study, several CAD software systems are applied
to construct a virtual design platform for the proposed
design method, permitting 3D scanning, 3D human
body modeling, 3D garment construction, 2D pattern
design, and 3D virtual try-on.
Human Solution ScanWorX. ScanWorX, developed by the
Human Solution Company, is used to control the body
scanning process and record the body scan result.
The body scan data obtained from the 3D scanning is
a set of points (point cloud) simulated as the shape of
the scanned object.
1723
Rapidform. Rapidform software was developed by
INUS Technology to edit 3D objects. Point clouds of
the scanning result will be simulated by the meshing
tool of Rapidform and a smooth body surface can be
obtained. Also, Rapidform ensures the 3D form can
be re-triangulated and the error holes, which are invari-
ably made as a result of scanning, will be filled. After
that, a unique surface modeled by many small facets is
obtained.
DesignConcept. DesignConcept, developed by Lectra, is
able to create points, lines, curves, and surfaces in a 3D
environment, which permits 3D garment prototyping
and flattening of corresponding 2D patterns.
Modaris. Modaris software, developed by Lectra, can be
used to create and edit 2D garment patterns and per-
form a virtual try-on process. The fabric parameters
of the proposed design can be modeled in the try-on
process at the same time.
Virtual 3D-to-2D garment block design process
In this study, a personalized garment block design
method is proposed. The proposed design method is
inspired by the general principle of the classic garment
block design method (Figure 1(a)) developed by Appel
and Stein: a garment surface framework can be obtained
from connecting different points of the human body with
a given allowance by different lines or curves following
the general shape of a garment.5
The proposed design method starts with 3D scanning.
The involved consumer is required to wear a bra and
underwear during the scanning procedure, ensuring
that the quality of the scanned result is good, especially
in the breast area. The scanning result will be fur-
ther modeled as a digitalized human body model. The
general process of human body modeling using the 3D
scan result has been explained in detail in previous
studies.5,16
The next procedure follows the conical principle
developed by Efrat, which is intended to establish 2D
co-ordinates for the crucial pattern shaping points. A
bodice is specified with several crucial shaping control-
ling points (SCPs). Then, the bodice is defined by sev-
eral triangular planes for both the front and back. The
bust and shoulder blades are the two prominent points,
which were used to generate these triangular planes.
Each triangle was formed by connecting two adjacent
perimeter points and the appropriate apex. Orthogonal
projections of the triangles are assembled to produce
front and back bodice patterns; proportional correc-
tions are made in the back shoulder area for the differ-
ence between the Euclidean distance calculated from
coordinates and the actual curvature of the back.
1724
Figure 1. The process of the proposed design method: (a) general principle of the classic garment block design method; (b) selected
anatomical landmarks for garment design; (c) definition of different shape controlling points, lines, and curves; (d) developing the
proposed garment block framework from the shape controlling points of the breasts and scapular points; (e) general principle of
defining the Normal Line of a curve. (Color online only.)
Firstly, a set of anatomical landmarks for garment
design (ALGDs; green points in Figure 1(b)) is selected
as part of the SCPs from the scanned body surface.
Eight ALGDs are selected in close connection
with the classic garment block design method,19 includ-
ing the waist, side, neck, armhole, and shoulder, as pre-
sented in Figure 1(b). Afterwards, the selected ALGDs
will be connected with lines or curves to create the gen-
eral framework of the virtual garment block. Lines are
marked with yellow and curves are marked with blue,
as presented in Figure 1(c). In the procedure of
developing the previous lines and curves, other SCPs
are developed from the human body surface with given
wearing ease (yellow points in Figures 1(b) and (c)).
The number and position of these SCPs also follow
the classic garment block generation method,19 includ-
ing one point on either side of the shoulder line and
three points on either side of the pattern perimeters.
After that, two prominent points, namely the apex
of the bust and the apex of the shoulder blade, are
used to generate the desired triangular planes, pre-
sented in a purple color in Figure 1(b). Twenty-four
SCPs are defined to determine the general framework
of the proposed garment block.
Using the pre-defined SCPs, a general framework of
the virtual garment block can be defined. A set of
curves is developed from the shape controlling points
of the apex of the bust and end with other SCPs to
define the 3D garment block surface using several tri-
angular planes. The shapes of these curves begin with
Textile Research Journal 88(15)
curves near the breast area, following the shape of the
breast, and end with straight lines when they turn to be
far away from the breast area. Finally, 28 triangular
planes are obtained, namely 13 for the front and
15 for the back panels. By adjusting the position of
the shape controlling points of the two breasts, an
ideal framework of the proposed virtual garment
block can be defined.
The distances between the SCPs of the breasts and
the real apex of the bust indicate the real wearing ease
of the breast area. As the design object is a garment
block, which contains very limited wearing ease, the
‘‘ideal position’’ of the shape controlling points of
the breast should be that which ensures the shortest
distance mentioned before, which consequently ensures
a smooth surface of the framework without any over-
lapping between the proposed garment block surface
and the human body.
Experimental details
In order to validate the proposed method, this research
presents case studies using the proposed method to
design personalized garment blocks for physically dis-
abled people of scoliosis type (PDPS). Scoliosis is
a physical condition that results in a curvature of
the spine.20 Scoliosis can cause atypical morphologies
characterized, in some cases, by an uncentered head,
uneven shoulders, ribs of different heights, uneven
hips, a shoulder blade that protrudes more than the
Hong et al.
other, one leg appearing shorter than the other, and the
body leaning to one side.21 More severe cases can pre-
sent a more severe outward appearance of these visible
symptoms. Because of the asymmetrical morphology
caused by scoliosis, the garment construction method
used for standard body shapes is not applicable for
PDPS. Due to the atypical morphology, PDPS were
chosen to be studied.
Twenty PDPS (C1, C2, . . . , C20) were involved in the
case studies. Only female PDPS are involved in the
experiments. These PDPS have serve scoliosis, accord-
ing to doctors’ diagnoses. They had a 50–55 degree
Cobb angle in adolescence.
The proposed case studies include four experiments,
as shown in Figure 2. In Experiment I, instrumental
testing of the fabric property is carried out, using the
Fabric Assurance by Simple Testing System (FAST).
Related fabric property information will be further
applied to simulate the virtual try-on of the garment
block. Also, real garment blocks will be produced
using real fabric to validate the design idea.
Experiment II explains how an application case of the
proposed design method is applied to obtain a perso-
nalized garment block for PDPS. Corresponding 2D
patterns will be generated and further applied to a 3D
virtual try-on procedure. Experiments III and IV are
validation experiments to evaluate the proposed
design method in terms of garment fitting. Several
KFMs are defined. These KFMs are further applied
to a normalized subjective visual evaluation in
Experiments III and IV. In Experiment III, subjective
visual evaluation is used by a group of designers to
obtain a normalized opinion of the garment fitting
Figure 2. Experiment design and working process of this study. KES: Kawabata Evaluation System.
1725
based on the experts’ knowledge and experience.
Several modifications will be carried out in conformity
with the designers’ recommendations. Corresponding
modified 2D patterns will be generated and utilized to
perform another 3D virtual try-on. In Experiment IV,
two real garments, using the primary garment patterns
and modified garment patterns, respectively, will be
produced using the proposed fabrics. Comparisons of
the pre-defined KFMs between the 3D simulation
results and real try-on results are carried out to validate
the proposed design method.
Experiment I
As the virtual fitting is strongly related to the fabric
properties, Experiment I is designed to obtain fabric
properties related to the virtual draping and try-on.
A piece of woven fabric of the greige cloth type,
which is widely used in clothing draping experiments,
is tested. Greige cloth is selected because it can repre-
sent the most popular fabrics used in apparel prod-
uct development. Normally, an objective measurement
of the sample fabric can be performed using the
FAST or the Kawabata Evaluation System (KES).22
Experimental data of the test results of both the
FAST and the KES can be uploaded to the Modaris
software to define fabrics in the simulation section of
Experiments II and III. In this study, the KES was
applied to the fabric property test. Compared with
the FAST, KESs are more specialized in fabric surface
performance testing, which is more applicable in this
research. By using the fabric properties presented in
Table 1, virtual fabric is simulated in Modaris software.
1726 Textile Research Journal 88(15)

3D method consisted of several design phases, which

Experiment II include 3D scanning (Section A in Figure 3), digitalized
In Experiment II, personalized garment block design 3D human body modeling (Sections B and C in
cases were performed using the proposed garment Figure 3), mechanical 3D garment block generation
design method. Figure 3 presents the design process (Sections D–G in Figure 3), interactive 3D garment
using the proposed design method and related software. surface (without fabric property) modeling, and 2D
As explained in the Methodology section, the proposed garment pattern design (Section H in Figure 3).
Using this process, a set of personalized garment
block patterns was generated. Openings and darts
Table 1. Fabric property of the sample fabric were given in this interactive process. The pattern
generation was an interactive process between the 3D
Fabric property criteria Fabric property virtual garment surface and the 2D garment block pat-
Raw material 100% Cotton (wrap and weft) tern. Several modifications of the generated 2D pattern
in a 2D environment using classic 2D pattern design
Weave structure 3/1/1/1/1/1 right-hand twill
knowledge were carried out to ensure that the flattened
Weight (g/m2) 156
2D could be further applied to the real simulation.
Thickness (mm) 0.596
These modifications included smoothing the serrated
Yarn fineness (tex ¼ g/km) Wrap 50 curves, correcting the length of corresponding curves
Weft 57 of pair patterns, and rectifying the position and
Yarn density (yarns/cm) Wrap 12 volume of darts.
Weft 13 Based on the flattened 2D garment block pattern
Bending (10–6Nm) Wrap 3.3735 obtained from the 3D garment surface, a digitalized
Weft 1.4024 human body model of the consumer, and the virtual
Shearing (Nm–1/ ) Wrap 1.471 fabric obtained from Experiment I, a 3D virtual try-
Weft 1.5985 on is carried out in Modaris software. Throughout
Tensile (N/m) Wrap 4.364 this operation, the 2D technical environment is lined
with a 3D virtual display environment. Based on the
Weft 13.0919
3D virtual try-on display, designers and consumers

Figure 3. Garment block generation process using the proposed design method, related software, and different working environ-
ments. 3D: three-dimensional; 2D: two-dimensional.
Hong et al.
Figure 4. Details of the design process for designing a persoanlized garment block. 3D: three-dimensional; 2D: two-dimensional.
were able to give feedback on the design result. The design
case of C1 is presented in Figure 4 to show the details of
the design procedure and the final result. Other design
cases are not presented due to the content limit.
Experiment III
Experiment III was proposed to evaluate the 3D virtual
try-on of the garment blocks of each PDPS. A group of
designers was invited to this section to make a group
decision and give feedback to the design result for each
design case.
1727
In order to identify the fit effect of the proposed
garment block, a set of normalized KFMs were
extracted, which constituted the common communica-
tion language between fashion designers, pattern
makers, and garment consumers. Five experienced
fashion designers were involved in the selection of
KFMs. Training sessions for the invited designers
about the purpose of this section were performed.
Instructions were given to designers to select a set of
measurements that meet both width and length require-
ments. Each trained designer generated an exhaustive
list of categories describing the garment fit performance
1728 Textile Research Journal 88(15)

Table 2. Key fit measurements (KFMs) describing Table 3. Evaluation results of all invited designers regarding the
the apparel fit performance garment block designed for C1

Categories Apparel KFMs FA1 FB1 FB2 FC1 FC2 FC3

FA: Fit in length FA1 Length D1 0 –1 –1 0 0 –1

FB: Fit in width FB1 Waist width D2 1 –1 –2 0 0 –2
FB2 Breast width D3 1 0 –1 0 –1 –1
D4 0 0 –1 0 0 –1
FC: Fit in details FC1 Shoulder fit D5 1 –1 –2 0 –1 –1
FC2 Neck fit
FC3 Arm hole fit
thought, which is full of uncertainty and imprecision.23
In this situation, fuzzy set theory was applied as a rele-
according to his/her professional knowledge. The most vant method for processing uncertain data obtained
relevant three categories describing the KFMs of from subjective evaluation.
the garment block were selected: ‘‘FA Fit in length’’, A triangular fuzzy number (TFN) M is usually
‘‘FB Fit in width,’’ and ‘‘FC Fit in details’’. Then, a denoted using n-tuple formalism as M ¼ (l, m, u) or
list of KFMs, characterizing garment fit in different M ¼ (l/m, m/u), where l, m, and u denote the smallest
categories, was generated by the designers using possible value, the most possible value, and the largest
their garment and pattern design knowledge. After possible value of a fuzzy event, respectively.24 Each set
that, a ‘‘round table’’ discussion among the panelists of TFNs has linear representations on its left- and
was performed to reduce redundant KFMs and those right-hand side. The membership function of a set of
irrelevant to fit. This step led to the generation of six TFNs can be defined as
normalized KFMs assessing the apparel fit perform- 8
ance (see Table 2). >
> 0, x 2 ½1, l
>
>
Subsequently, a scale of five evaluation scores, ran- >
> xl
< , x 2 ½l, m
ging from –2 to 2 (–2, –1, 0, 1, 2), was defined to m l
m ðxÞ ¼ ð1Þ
describe KFMs. ‘‘–’’ means that the garment is tight >
> x  u
>
> , x 2 ½m, u
or small related to the body shape, while ‘‘+’’ goes in >
> mu
:
the opposite direction (large or loose) and 0 means a 0, x 2 ½u, þ 1
perfect fit on the wearer.
Using these KFMs, group decision-makings were If M1 ¼ ðl1 , m1 , u1 Þ and M2 ¼ ðl2 , m2 , u2 Þ are two
performed in order to assess the fit effect of the 3D TFNs, the operation laws between them can be defined as
virtual try-on result obtained in Experiment II.
Firstly, five experienced designers (D1, D2, D3, D4, M1  M2 ¼ ðl1 þ l2 , m1 þ m2 , u1 þ u2 Þ ð2Þ
and D5) were invited to evaluate the 3D virtual try-on
result using the proposed KFMs. The proposed 3D M1  M2 ¼ ðl1 l2 , m1 m2 , u1 u2 Þ ð3Þ
virtual try-on result was displayed on a screen in
front of the invited designers. Each of the designers l  M1 ¼ ðll1 , lm1 , lu1 Þ, l 2 R ð4Þ
was assigned to evaluate the 3D virtual try-on result  
independently without previous discussion with other 1 1 1
ðl1 , m1 , u1 Þ1 ¼ , , ð5Þ
designers. Designers were free to operate the computer u1 m1 l1
to observe the 3D model under all aspects.
Taking C1 as an example, each of the designers was where  denotes the extended summation of two TFNs
requested to record the fit effect in terms of KFMs pre- and  denotes the extended multiplication.
sented in Table 2, using the pre-defined evaluation Euclidean distance can be used to describe the dis-
score. Table 3 presents the evaluation result of all tance between two TFNs.25 The distance between M1 ¼
invited designers about the fit effect of the 3D virtual ðl1 , m1 , u1 Þ and M2 ¼ ðl2 , m2 , u2 Þ 26 can be expressed as
try-on result of C1.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Experiments designed for the fit effect assessment are 1 
based on subjective visual evaluation. In this procedure, ~ mÞ
d ðn, ~ ¼ ðl1  l2 Þ2 þðm1  m2 Þ2 þðu1  u2 Þ2 ð6Þ
3
uncertain and imprecise linguistic expressions are often
used by both designers and consumers for evaluating Using fuzzy set tools, a set of linguistic terms,
the garment block. The fuzzy set, as an intelligent tech- describing the evaluation criteria, can be quantified in
nique, was developed to handle the vagueness of human the universe of respective and discourse membership
Hong et al. 1729

Figure 5. The fuzzy linguistic rating scale and related triangular fuzzy numbers.

Table 4. Evaluation scores, corresponding linguistic terms and Using the same calculation process, the perception
triangular fuzzy numbers (TFNs) of all the designers can be aggregated in terms of dif-
ferent KFMs, as presented in Table 5.
Evaluation
scores Linguistic terms TFNs After that, in order to investigate the fit effect of
different garment block positions regarding various
2 Very loose/big (VL) (2.5,3,3.5) KFMs, the distance of all the aggregated TFNs are
1 A little loose/big (AL) (2,2.5,3) measured to the ‘‘Perfect’’ condition, whose corres-
0 Perfect (P) (1.5,2,2.5) ponding TFN is (1.5,2,2.5). When the distance is
–1 A little tight/small (AT) (1,1.5,2) shorter, the satisfaction is higher. For example, using
–2 Very tight/small (VT) (0.5,1,1.5) equation (6), the fit effect of FA1 can be calculated as
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 
function (Figure 5). TFNs, as a classic fuzzy set tool, ð1:8  1:5Þ2 þ ð2:3  2Þ2 þ ð2:8  2:5Þ2 ¼ 0:17
3
are used to quantify the utilized linguistic terms in this
research. Each score of the scale is defined by a corres- Similarly, all the aggregated TFNs and correspond-
ponding linguistic term. In order to quantify these lin- ing distance to the ‘‘Perfect’’ condition can be formu-
guistic terms, corresponding TFNs are also given, as lated as presented in Table 5
presented in Table 4. Based on the result obtained, a group discussion was
Then, the aggregation operation of the quantized organized again to decide a pattern medication plan.
evaluation result is performed using the following Then, the garment patterns were modified in a 2D
equation environment according to the professional knowledge
of pattern designers. The modified pattern was simu-
!
1X l
1X l
1X l lated in a 3D virtual environment again to be analyzed
aij ¼ aijh t1 , aijh t2 , aijh t3 ð7Þ by the designers. The same evaluation procedure that
m j¼1 m j¼1 m j¼1
was applied at the previous stage was carried out again.
The sequence of Design – Display – Evaluation –
where t1, t2, and t3 correspond to the value of TFNs, Adjustment can be performed repeatedly until a satisfy-
which are taken according to Table 4. Using this ing design solution is obtained. Figure 6 presents the
method, the evaluation results of the designer can be modified garment block patterns and corresponding
quantified and aggregated into a group decision related virtual try-on result for C1.
to the fit effect of the proposed garment block displayed
in the 3D virtual try-on. Experiment IV
For example, for the group perception of the invited
designers (D1, D2, D3, D4, and D5) in terms of FA1 of the Experiment IV was designed to validate the proposed
garment block design for C1, can be formulated as a design process. Real samples for both the initial design
new TFN using equation (7) result and the modified design result were simulated
using the proposed fabric in Experiment I and different
  garment block patterns in Experiments II and III.
1:5  2 þ 2  3 2  2 þ 2:5  3 2:5  2 þ 3  3
, , Figure 7 presents the two real samples of the design
5 5 5
case of C1. Another group of five experienced designers
¼ ð1:8, 2:3, 2:8Þ were also invited to evaluate the two real samples tried
1730 Textile Research Journal 88(15)

Table 5. Aggregated evaluation result and corresponding distance to the ‘‘Perfect’’ condition of the design case of C1

FA1 FB1 FB2 FC1 FC2 FC3

Aggregated evaluation result (1.8,2.3,2.8) (1,1.5,2) (0.7,1.2,1.7) (1.5,2,2.5) (1.4,1.9,2.4) (0.8,1.3,1.8)

Distance to the ‘‘Perfect’’ condition 0,17 0,17 0,40 0,00 0,12 0,35

Figure 6. The modified garment block patterns and corresponding virtual try-on result for C1.

Figure 7. Different views of two sets of real samples of C1 produced by initial and modified garment block patterns, and the
proposed woven fabric of the greige cloth type.

on by C1, according to KFMs presented in Table 2, other designers. During the evaluation session,
using the evaluation scores presented in Table 4. Each designers were free to observe the consumer with the
of the designers was assigned to evaluate the 3D virtual two different pieces of garment block from various
try-on result one by one, avoiding any discussions with angles. Each of the designers was requested to record
Hong et al. 1731

the fit effect in terms of the KFMs in Table 2, using the formulated. Using equation (6), the dissimilarity between
pre-defined evaluation score. After the evaluation of ñ (n1, n2, n3) and m~ (m1, m2, m3) can be denoted as
each designer, there was a rest period for the PDPS.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi
1 2 2 2
3 ð n 1  m 1 Þ þ ð n 2  m 2 Þ þ ð n 3  m 3 Þ
~ mÞ
d ðn, ~ ¼ ð8Þ
Results and discussion 10

Results and discussion of the design case of C1 Correspondingly, the similarity between n~ and m~
Evaluation results for real samples of Experiment IV could be defined as
were collected and aggregated using equation (7), as
presented in Table 6. According to the result presented ~ m~ Þ ¼ ð1  dðn,
sðn, ~ m~ ÞÞ  100%
in Table 6, distances of the quantified designers’ per- 0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 ffi1
1 2 2 2
ception about the fit effect between different garment 3 ðn 1  m 1 Þ þ ðn 2  m 2 Þ þ ðn 3  m 3 Þ
¼ @1  A
positions on the real garment block sample regarding 10
different KFMs and the ‘‘Perfection’’ condition are all
very short; thus, it can be concluded that the satisfac-  100%
tion of the real garment block sample, produced using ð9Þ
the modified garment block patterns, is very high in
general. Using equation (9), the similarity degree of the
This indicates that, through a subjective visual evalu- aggregated evaluation results between virtual and real
ation performed by the invited designers regarding the garment try-on results using the initial garment block
proposed KFMs, the design effect can be effectively pattern, regarding different KFMs, can be obtained.
analyzed to generate a knowledge-oriented pattern For example, concerning FB1 Waist width, the similarity
modification section, which can fully ensure the satis- degree of the aggregated evaluation results between the
faction of the final product. As a knowledge-based pro- virtual result (1,1.5,2) and the real garment try-on result
cess, the knowledge and experience of the designers can (1.2,1.7,2.2) can be calculated as
be largely extracted and applied to support the design
0 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 1
process and enhance the desired garment design effect. 1
ð 1  1:2 Þ 2
þ ð 1:5  1:7 Þ 2
þ ð 2  2:2 Þ 2
Also, in order to validate if the proposed design @1  3 A
method is a normalized method that can be generally 10
applied to provide a design solution for consumers with
 100% ¼ 88:45%
atypical morphology, comparisons of the similarity
degrees between the evaluation results of virtual samples
and real samples using the initial garment block patterns All the similarity degrees were calculated in the same
were performed. The similarity degree is a score to be way and the results are presented in Figure 7.
uniformly distributed in the interval of [0,1].27 In order From Figure 8 it can be seen that the highest simi-
to obtain the similarity degree, the following equation was larity degree is 100% (FA1 Length and FC1 Shoulder fit),

Table 6. Aggregated evaluation result for garment samples of C1, produced by initial and modified garment block patterns and
corresponding distances to the ‘‘Perfect’’ condition

Aggregated evaluation result FA1 FB1 FB2 FC1 FC2 FC3

Real sample of initial garment (1.8,2.3,2.8) (1.2,1.7,2.2) (0.8,1.3,1.8) (1.5,2,2.5) (1.3,1.8,2.3) (0.9,1.4,1.9)
block patterns
Real sample of modified (1.33,1.83,2.33) (1.5,2,2.5) (1.67,2.17,2.67) (1.5,2,2.5) (1.33,1.83,2.33) (1.17,1.67,2.17)
garment block patterns

Distance to the ‘‘Perfect’’ FA1 FB1 FB2 FC1 FC2 FC3

condition
Real sample of initial 0,17 0,29 0,46 0,00 0,06 0,40
garment block patterns
Real sample of modified 0.10 0.00 0.10 0.00 0.10 0.19
garment block patterns
1732
Figure 8. The similarity degree of the perception of designers between virtual and real samples using the initial garment block
pattern of the design case of C1.
Figure 9. Average similarity degree between virtual and real samples using the initial garment block pattern regarding different key fit
measurements of all design cases based on the perception of designers.
most of the similarity degrees are higher than 94%
(FA1 Length, FB2 Breast width, FC1 Shoulder fit, FC2
Neck fit, and FC3 Arm hole fit), and the lowest is 88%
(FB1 Waist width). It can be concluded that the general
similarity degrees of the aggregated evaluation results
between virtual and real garment try-on results, using
the initial garment block pattern, are at a high level,
which indicates that different groups of designers can
have a common perception with the real and virtual
perception, and current virtual software can simulate
virtual products with a high fidelity.
Results and discussion of all the design cases
Figure 9 presents the average similarity degree between
virtual and real samples of all design cases, evaluated
by designers. From Figure 9, the minimum average
similarity degree is 84.31% (C11) and the maximum is
96.53% (C18). The average of all these similarity
degrees is 91.65%. This indicates that the similarity
degree between virtual and real samples maintains a
high level. Even though the visual evaluation is subject-
ive, the proposed evaluation method, using evaluation
criteria realized by a set of KFMs, is a normalized
Textile Research Journal 88(15)
method, which can generally be applied to provide
a design solution for consumers with atypical morph-
ology. The application of subjective visual evaluation,
according to the proposed KFMs, design knowledge,
and experience of the designers can be extracted to
support the design process and ensure a stable precision
of the result. The experience and professional know-
ledge of the designers can be applied to evaluate the
design effect and easily obtain a common agreement
without conflict.
Several experienced designers were involved in the
proposed fit effect evaluation method; the sensitivity
of the subjective data, obtained by subjective visual
evaluation, can be largely reduced. The fuzzy logic-
based group decision-making method works properly
in knowledge extraction with stable quantified evalu-
ation results.
The proposed sequence of Design – Display –
Evaluation – Adjustment can be effectively applied to
the personalized garment design, especially for atypical
morphology with an ensured stability and precision of
the results. Also, virtual reality technology is able to
simulate garment products in a virtual environment
with a high similarity degree of the real garment
Hong et al.
product, which indicates that the proposed digitalized
3D simulation-based design process can ensure a high
accuracy.
Conclusion
In this research, a personalized ready-to-wear garment
product design process is proposed to provide a design
solution for consumers with atypical morphology.
The proposed design solution starts with designing a
personalized garment block and then this is extended
to garment patterns in order to obtain the desired
ready-to-wear garment style. The proposed garment
block is designed through a virtual 3D-to-2D design
process and the extension operation is realized using
the classic 2D pattern design method, avoiding compli-
cated operations of simulating a 3D garment directly in
the virtual environment. The proposed design process
can be fully digitalized, which ensures the involvement
of the consumer in the whole design process.
Two experiments were designed to present the pro-
posed sequence of Design – Display – Evaluation –
Adjustment by explaining a design case for designing
a personalized garment block for PDPS, which is tech-
nically accomplished through the virtual 3D-to-2D
design method and 3D virtual try-on. Another two
experiments were designed to validate the proposed
design process. The experimental results show that the
proposed design process is a normalized design method,
which is able to efficiently provide technical solutions
for garment design aiming at consumers with atypical
morphology. As a knowledge-supported design pro-
cess, the design knowledge and experience of designers
can be fully extracted to enhance the satisfaction of the
final product. The proposed subjective visual evalu-
ation method using a set of normalized KFMs is able
to obtain a general outcome of the fit effect. Fuzzy logic
can quantify the results, ensuring a common stable
evaluation. Using the proposed design sequence,
experts’ design ideas and principles of designing perso-
nalized garment products can be validated within a very
short time.
Acknowledgement
Special thanks to the Lectra Company for support for the
software.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with
respect to the research, authorship and/or publication of this
article.
Funding
The authors disclosed receipt of the following financial sup-
port for the research, authorship, and/or publication of this
1733
article: This work was supported in part by the SMDTex
Project, funded by the European Erasmus Mundus Program.
References
1. Agarwal G, Koehl L, Zeng X, et al. An intelligent system
for supporting design of fashion oriented personalized
fabric products. Ind J Fibre Text Res 2009; 34: 258–266.
2. Du X, Jiao J and Tseng MM. Understanding customer
satisfaction in product customization. Int J Adv Manuf
Technol 2006; 31: 396–406.
3. Kim S, Jeong Y, Lee Y, et al. 3D pattern development of
tight-fitting dress for an asymmetrical female manikin.
Fibers Polym 2010; 11: 142–146.
4. Montagna G, Carvalho H, Catarino AP, et al.
Ergonomic garment design for a seamless instrumented
swimsuit. In: Koglbauer L (ed.) 4th international confer-
ence on applied human factors and ergonomics (AHFE),
Hilton San Francisco Union Square, California, 21–25
July 2012, pp.1204–1213. New York: USA Publishing,
127.
5. Hong Y, Zeng X, Bruniaux P, et al. Interactive virtual
try-on based three-dimensional garment block design for
disabled people of scoliosis type. Text Res J 2017; 87:
1261–1274.
6. Hong Y, Bruniaux P, Zeng X, et al. virtual reality based
collaborative design method for designing customized
garment of disabled people with scoliosis. Int J Clothing
Sci Technol 29: 226–237.
7. Hernández N. Tailoring the unique figure. Göteborgs:
Göteborgs universitet, 2000.
8. Sayem ASM, Kennon R and Clarke N. 3D CAD systems
for the clothing industry. Int J Fashion Des Technol Educ
2010; 3: 45–53.
9. Wang CCL, Wang Y and Yuen MMF. Feature based 3D
garment design through 2D sketches. Comput Aid Des
2003; 35: 659–672.
10. Liu Y-J, Zhang D-L and Yuen MM. A survey on CAD
methods in 3D garment design. Comput Ind 2010; 61:
576–593.
11. Chen X, Tao X, Zeng X, et al. Control and optimization
of human perception on virtual garment products
by learning from experimental data. Knowl Based Syst
2015; 87: 92–101.
12. Hong Y, Curteza A, Zeng X, et al. Sensory evaluation
based fuzzy AHP approach for material selection in cus-
tomized garment design and development process.
In: Ashton A (ed.) Book of abstracts, London: IOP
Publishing, 2016, p.82.
13. Hong P, Doll WJ, Nahm AY, et al. Knowledge sharing in
integrated product development. Eur J Innovat Manag
2004; 7: 102–112.
14. Zhang D, Wang J and Yang Y. Design 3D garments for
scanned human bodies. J Mech Sci Technol 2014; 28:
2479–2487.
15. Han HS, Kim JY, Kim SM, et al. The development of an
automatic pattern-making system for made-to-measure
clothing. Fibers Polym 2014; 15: 422–425.
1734
16. Thomassey S and Bruniaux P. A template of ease allow-
ance for garments based on a 3D reverse methodology.
Int J Ind Ergon 2013; 43: 406–416.
17. Huang H, Mok P, Kwok Y, et al. Block pattern gener-
ation: From parameterizing human bodies to fit feature-
aligned and flattenable 3D garments. Comput Ind 2012;
63: 680–691.
18. Shin K. Patternmaking for the underwired bra: new dir-
ections. J Text Inst 2007; 98: 301–318.
19. Efrat S. The development of a method for generating pat-
terns for garments that conform to the shape of the human
body. De Montfort University, Leicester, 1982.
20. Stokes IA. Three-dimensional terminology of spinal
deformity: a report presented to the Scoliosis Research
Society by The Scoliosis Research Society Working
Group on 3-D terminology of spinal deformity. Spine
1994; 19: 236–248.
21. Thong W, Parent S, Wu J, et al. Three-dimensional
morphology study of surgical adolescent idiopathic scoli-
osis patient from encoded geometric models. Eur Spine J
2016; 25: 1–10.
Textile Research Journal 88(15)
22. Yick K-L, Cheng K, Dhingra R, et al. Comparison of
mechanical properties of shirting materials measured on
the KES-F and FAST instruments. Text Res J 1996; 66:
622–633.
23. Dubois D and Prade H. On the use of aggregation oper-
ations in information fusion processes. Fuzzy Sets Syst
2004; 142: 143–161.
24. Zeng X, Ding Y and Koehl L. A 2-tuple fuzzy linguistic
model for sensory fabric hand evaluation. In: Xianyi Z
(ed.) Intelligent sensory evaluation. Berlin: Springer, 2004,
pp.217–234.
25. Zhu Y, Zeng X, Koehl L, et al. A general methodology
for analyzing fashion oriented textile products using sen-
sory evaluation. Food Qual Pref 2010; 21: 1068–1076.
26. Zhu K-J, Jing Y and Chang D-Y. A discussion on extent
analysis method and applications of fuzzy AHP. EurJ
Oper Res 1999; 116: 450–456.
27. Hong DH and Kim C. A note on similarity measures
between vague sets and between elements. Inform Sci
1999; 115: 83–96.