International Journal of Industrial Ergonomics 72 (2019) 212–221

Contents lists available at ScienceDirect

International Journal of Industrial Ergonomics

journal homepage: www.elsevier.com/locate/ergon

Parametric design of garment pattern based on body dimensions T

a,b,d,∗ c d d d
Kaixuan Liu , Chun Zhu , Xuyuan Tao , Pascal Bruniaux , Xianyi Zeng
a
Apparel and Art Design College, Xi'an Polytechnic University, Xi'an, 710048, China
b
Shanxi Key Laboratory of Intelligent Clothing Design, Xi'an Polytechnic University, Xian, 710048, China
c
Human Resources Office, Xi'an Polytechnic University, Xi'an, 710048, China
d
GEMTEX Laboratory, ENSAIT, 59100, Roubaix, France

A R T I C LE I N FO A B S T R A C T

Keywords: Garment pattern-making is one of the most important parts in fashion design and production. However, the
Body dimensions traditional pattern-making is an experience based work and very time-consuming. In this paper, we proposed a
Pattern-making parametric design method of garment pattern based on body dimensions. Based on this method, we constructed
Parametric design a jeans' pattern recommendation system. The input items of the proposed system are three geometric constraint
CAD
parameters (jean silhouette type, length and waist height) and three-dimensional constraint parameters (human
Knowledge modeling
Anthropometric measurement
body stature, waist girth and hip girth); the output of the proposed system are jeans' patterns. Also, four ad-
justable parameters (jeans' length, waist height, knee and leg opening) are designed to adjust patterns generated
by the proposed system. If the jeans' pattern is not satisfying after virtual or real try-on, the adjustable input
parameters of the proposed system can be applied for adjustment until the patterns are acceptable. Our proposed
system can combine traditional pattern-making methods to generate jeans’ patterns automatically and rapidly,
hence improving pattern-making efficiency significantly.

1. Introduction consuming and limits the improvement of production efficiency; the

other is that this technique is difficult to spread. For a beginner, it takes
In the clothing industry, customers' individual needs have become a long time to study, three to five years just to get in the door as a
increasingly emphasized and changing over time. This is partly due to pattern maker, and at least ten years of mastery. In order to enhance the
the development of economies, and societies, and the advance from pattern-making speed, pattern makers develop patterns for a standard
cottage industry type, to Mass production and Made to Measure (MTM), body type size and use grading technology to obtain other body type
to mass customization (Kamalha et al., 2013; Satam et al., 2011; Zhang sizes (Cho et al., 2006). Due to the differences in the body measure-
et al., 2012). However, the sophistication in consumers’ demands and ments, this technology doesn't fit the body well. This is why consumers
fast fashion result in increased difficulty in clothing making. Mass usually complain that they can't find firm-fitting garments (Liu et al.,
customization, a hybrid of Mass production and Made to Measure (Yang 2017a; Liu et al., 2017c). Currently, MTM can solve the fitting issues;
et al., 2007), has become a trend in clothing industry; integrating the nevertheless, this method results in high manpower cost.
efficiency and personalized service excellence of the two production In order to solve fitting issues, many scholars used 3D-to-2D flatten
modes (Cho et al., 2006; Li et al., 2013; Yang et al., 2007). The emer- technology to develop garment patterns (Liu et al., 2016c; Liu et al.,
gence of mass customization raises new demands for pattern-making; 2016a; Liu et al., 2018a). However, this method has four shortages: the
faster and more fitting. However, faster and more fitting are contra- first is that this method is only suitable for simple styles, and not sui-
dictory targets to be achieved by traditional pattern-making methods. table for complicated styles; secondly, the ease allowances between the
Pattern-making is one of the most critical and technical elements in human body and clothing are not determined accurately; the third is
clothing design and manufacturing (Liu et al., 2016c; Liu et al., 2018a). that the fabric properties are rarely considered in the flattening process;
The speed of pattern-making is crucial for efficiency gains. At present, the fourth is that this method is quite time-consuming and not suitable
there are two methods to make garment pattern: hand making and for industrial production. As such, there's been less progress on this
computer-aided making. Both the two methods require sophisticated approach and rarely implemented into practical production.
skills. Moreover, the traditional pattern-making method has two Some scholars proposed parametric pattern design methods to de-
shortcomings: one is that the pattern-making process is quite time- velop garment patterns and fashion sketches rapidly (Kim, 2012; Liu

∗
Corresponding author. Apparel and Art Design College, Xi'an Polytechnic University, Xi'an, 710048, China.
E-mail addresses: [email protected], [email protected] (K. Liu).

https://doi.org/10.1016/j.ergon.2019.05.012
Received 28 January 2019; Received in revised form 28 April 2019; Accepted 28 May 2019
Available online 04 June 2019
0169-8141/ © 2019 Elsevier B.V. All rights reserved.
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

et al., 2019; Liu et al., 2018b). However, there has been less progress on generate garment patterns automatically.
this aspect and rarely implemented into practical production. The two
main reasons are: firstly, these studies are only for a fixed clothing style, 2.2. Basic function of jeans’ pattern recommendation system
rather than a series of clothing styles. However, one style cannot meet
the market demand. The second is that they lack a feedback in the The basic functions of jeans' pattern recommendation system are
parametric process. For example, if a client is not satisfied with the described in Fig. 1. Its basic functions are given below. A user inputs his
garment, there is no approach to adjust the result to meet the client's body dimensions and selects jean style. Using our parametric model, the
different requirements. proposed system can generate garment patterns automatically. Next is
In order to solve the issues in parametric pattern design, we pro- to adjust the 3D human body model according to the user's body di-
posed a novel parametric pattern-making model. The proposed method mensions. Then, the patterns are assembled on the adjusted virtual
can generate patterns rapidly and automatically according to different avatar using virtual try-on technology. The user then evaluates the
jean styles. If a user is dissatisfied by the result, it has a feedback to virtual garment. If the result is satisfactory, then the garment is pro-
modify the pattern until the pattern fits the body well. The construction duced; otherwise, the patterns are adjusted using adjustable parameters
of this model mainly involves three steps. Firstly, we carried out a until it meets requirements.
program of anthropometric measurements. We randomly selected 116
young women aged 20–30 years from the Northeast area of China.
3. Methodology
Secondly, we applied factor analysis, regression analysis and design
knowledge to acquire the linear relations between human body stature,
3.1. Acquisition of initial input parameters and adjustable input parameters
hip girth, waist girth and other body dimensions. Finally, a series of
parametric models was built according to these linear equations.
3.1.1. Graphics constraints of the jeans’ patterns
Finally, some parameters were set in the models to modify patterns to
From the perspective of graphics, graphics constraints are divided
fit body according to client's feedback information.
into two categories: dimensional and geometric constraint (Xiu et al.,
We selected jeans as the research focus mainly due to the universal
2011). As shown in Fig. 2 (a), the patterns of bell-bottom jeans and
prevalence of jeans use all over the word. Sure, our method is not
pencil jeans have different geometric shapes, so are geometrically
limited to jeans, it can be applied in other fashion styles. This article
constrained. The two pairs of bell-bottom jeans in Fig. 2 (b) being
was organized as follows: section 2 presented the general scheme of
different in form of dimensions; therefore, are dimensionally con-
jeans’ pattern recommendation system; section 3 discussed how to
strained. These two types of constraints can finally fix the shape of
construct the proposed system using the parametric design method;
jeans' patterns. It is necessary to establish geometric constraint para-
section 4 introduced the application and validation of the proposed
meters and dimensional constraint parameters for jeans' patterns. For
system; finally, some conclusions and possible further works were
jeans’ patterns, the geometric constraint is style, and dimensional
presented in section 5.
constraints are pattern dimensions.

2. General scheme of jeans’ pattern recommendation system

3.1.2. Definition of geometric constraint parameters
2.1. Construction of jeans’ pattern recommendation system Two steps of questionnaire surveys were carried out to extract jean
classification criteria (geometric constraint parameters). In the first
The construction of jeans’ pattern recommendation system is de- stage, 20 fashion designers were surveyed to collect influencing factors
scribed in Fig. 1. It consists of four parts: initial input part, adjustable for jeans styles. The 20 fashion designers are familiar with the design of
input part, parametric model part and output part. Its basic steps are as women's jeans in the 20–30 age group. The first-stage survey results
follows: showed that jeans style mainly depends on three factors, i.e., silhouette,
length and waist height type (Fig. 3). We defined the three factors as

• Designing the initial input parameters of the proposed system. three parameters. Thereafter, based on the results, the three parameters

• Designing the adjustable input parameters of the proposed system. were further subdivided through a second survey of 20 fashion de-

• Modeling the relations between jeans' patterns and initial input signers. The subdivision results in Fig. 3 were selected as the initial
input items of the proposed recommendation system in Fig. 1.
parameters.
• Modeling the relations between jeans' patterns and adjustable input 3.1.3. Definition of dimensional constraint parameters
parameters.
• Developing an application according to the above models to If we want to make patterns for jeans, every jean's part dimension
should be established. However, jeans' patterns consist of many curves.

Fig. 1. General scheme of jeans' pattern recommendation system.

213
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 2. Geometric constraint and dimensional constraint.

Each curve's dimension is a kind of dimensional constraint. Too many knee girth, waist girth, hip girth and crotch width were selected as the
dimensional constraints complicate pattern-making. In order to simplify adjustable parameters to construct the proposed system.
pattern-making, we should use a few key dimensional constraint
parameters to represent other secondary dimensional constraints 3.2. Modeling the relations between geometric constraint parameters and
parameters. Research has showed that all dimensions of pants in the jeans dimensions
height direction can be represented by stature and in the girth direc-
tion, can be represented by hip girth and waist girth (Liu et al., 2017b; In the previous section, we designed the input parameters for jeans'
Liu et al., 2016b). Thus, the stature, hip girth and waist girth were pattern recommendation system. In this section, we model the relations
selected as the initial input items of the proposed recommendation between input parameters and jeans’ pattern dimensions.
system in Fig. 1.

3.2.1. Modeling the relations between jean waist parameter and jeans’
3.1.4. Acquisition of adjustable input parameters of jeans’ pattern pattern dimensions
recommendation system As shown in Fig. 4, the high waist, normal waist and low waist jeans'
Currently, the parametric design of garment patterns only focus on waist lines correspond with the human body waist line, lower waist line
how to make patterns rapidly (Xiu et al., 2011). However, there is no and abdomen line respectively. The lower waist line is about at the
means to modify the pattern if the consumer is not satisfied with the middle of the human body waist line and abdomen line. Different jeans'
jeans. For example, if the jean is too long or the leg opening is too wide, waist types mainly influence the parts above crotch depth line and do
how to modify? In this research, adjustable input parameters are de- not influence the parts below it (Fig. 4). Thus, drop of waist-to-hip and
signed to adjust patterns, which do not meet satisfaction after try-on, by crotch depth are involved in waist design; and they are two key di-
the wearer. Through questionnaire survey, we established that con- mensions for pattern-making of bottoms. According to the relations
sumers' dissatisfaction usually focuses on two aspects: appearance dis- between jeans and the human body in Fig. 4, jeans’ drop of waist-to-hip
satisfaction and comfort dissatisfaction. The appearance dissatisfaction and crotch depth are defined as follows:
contains four factors: jeans’ length, waist height type, leg opening girth
and knee girth; and the comfort dissatisfaction contains four factors: DWHj = DWHh + fDWA (pwaist ) (1)
waist girth, hip girth and crotch width. Thus, these seven factors were
selected as seven adjustable input parameters for pattern adjustment in CDj = CDh + fDWA (pwaist ) (2)
case of need.
Finally, through the above analysis, three geometric constraint ⎧ 0 pwaist = high waist
parameters (jean waist height type, silhouette and length type) and fDWA (pwaist ) = − DWAh /2 pwaist = normal waist
⎨
three-dimensional constraint parameters (human stature, hip girth and
⎩ − DWAh pwaist = low waist (3)
waist girth) were selected as the initial input parameters to construct
the proposed system. Jeans’ length, waist height, leg opening girth, where,

Fig. 3. Jean style classification.

214
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 4. Jean waist design. (WHh is human body waist height; CDh is human body crotch depth; DWHh is human body drop of waist-to-hip; DWAh is human body drop
of waist-to-abdomen.)

DWHj is jeans' drop of waist-to-hip; ⎧ 0 psilhouette = "V "

DWHh is human body's drop of waist-to-hip; ⎪ 0 psilhouette = "H "
CDj is jeans' crotch depth; fK (psilhouette ) =
⎨ depend on the value of LOj psilhouette = "X "
CDh is human body's crotch depth; ⎪
depend on the value of LOj psilhouette = "A" (6)
DWAh is human body's drop of waist-to-abdomen; ⎩
pwaist is the parameter of waist height; it includes three values: high,
normal and low waist (refer to Fig. 4).
⎧− α psilhouette = "V "
⎪ 0 psilhouette = "H "
Definitions (1), (2) and (3) are used for the parametric design of fLO (psilhouette ) =
jean waist types. ⎨ β psilhouette = "X "
⎪
⎩ γ psilhouette = "A" (7)
3.2.2. Modeling the relations between jeans' silhouette parameter and jeans’
Where,
pattern dimensions
As shown in Fig. 5, the area between waist line and crotch depth
Kj is jeans' knee girth;
line is the fit-zone; and the area below the crotch depth line is the
LOj is jeans' leg opening girth;
design-zone. Jeans should fit the body well in the fit-zone; meanwhile,
Hh is human body hip girth;
designers can design various jeans in this zone. Thus, jean styles'
eh is Ease allowance at hip; its value depends on the design re-
characteristics mainly depend on the part below crotch depth line.
quirement.
Different jeans' silhouettes mainly influence the parts below crotch
a is a coefficient; its value depends on the ratio of knee -to-hip girth;
depth line and do not influence the parts above it (Fig. 5). Thus, knee
α, β and γ are positive constants, whose values are given by the
girth and leg opening girth are involved in silhouette design; and they
designer.
are two key dimensions for pattern-making of bottoms. According to
psilhouette is a silhouette parameter, which includes four values: “A”
the relations among the four silhouette types in Fig. 5, jeans’ knee girth
type, “V” type, “X” type and “A” type (refer to Fig. 5);
and leg opening girth are defined as follows:

Kj = a (Hh + eh) + fK (psilhouette ) (4) Definitions (4), (5), (6) and (7) are used for jeans’ silhouette para-
metric design.
LOj = a (Hh + eh) + fLO (psilhouette ) (5)

Fig. 5. Jeans' silhouette design.

215
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 6. Jeans' length design. (Note: The jeans refer to height waist type's jeans. WHh is human body Waist Height. LLh is human body Leg Length. KHh is human body
Knee Height.).

3.2.3. Modeling the relations between jeans' length parameters and jeans’ 3.3. Modeling the relations between dimensional constraint parameters and
dimensions human body dimensions
As shown in Fig. 6, for the hot length style, the length line in the
style figure is usually d cm below crotch depth line. The fifth jean's In our previous analysis, jeans' shape is constrained by geometric
length line overlaps with the human body's knee line. The long jean's constraints and dimensions constraints; and we modeled the relations
length line overlaps with the human body heel line. The third jean's between geometric constraint parameters and jeans' pattern dimen-
length line is in the middle of hot jean's length line and the fifth jean's sions. In the following sections, we model the relations between di-
length. The seventh and ninth jean's lines are the trisections between mensional constraint parameters and human body dimensions. Thus, all
human body knee line and heel line. According to the relations between dimensions of jeans’ patterns can be represented using several key
jeans and human body in Fig. 6, jeans' lengths are defined as follows: human body dimensional parameters (hip, waist and stature).
JL = CDj + fJL (plength ) (8) 3.3.1. Relations between human body and jeans’ pattern
where, As shown in Fig. 7, jeans' patterns' dimensions have a one-to-one
relation with human body dimensions in the height direction; jeans’
JL is jeans' length; pattern dimensions are equal to corresponding human body girths plus
CDj is jeans' crotch depth; ease allowance in girth direction. For example, human knee height
plength is the parameter of jeans' length design; it includes six values: corresponds to jean knee height (Fig. 7 (a)); jean hip girth is equal to
hot, third, fifth, seventh, ninth and long (refer to Fig. 6). human body hip girth plus ease allowance (Fig. 7 (b)). Thus, we can
measure human body dimensions, and use these dimensions to make
Then, substituting Eq. (3) into Eq. (9), we obtain Eq. (3) as follows: patterns. However, it is very time-consuming to measure all the related-
pattern-making dimensions. In order to simplify pattern-making pro-
JL = DWHh + fDWA (pwaist ) + fJL (plength ) (9) cess, it is necessary to use several key dimensions to represent other
secondary dimensions. In our previous analysis, we selected stature and
⎧ d plength = hot hip as the key parameters to represent other height direction dimen-
⎪ d + (LLh − d − KHh)/2 plength = third sions and girth direction dimensions respectively. As linear models are
⎪ widely applied in the garment pattern-making (Xiu et al., 2011), we
⎪ LLh − KHh plength = fifth
fJL (plength ) = selected a linear model to construct the relations between stature, hip
⎨ LLh − 2/3 × KHh plength = seventh girth and other human body dimensions in the following sections.
⎪
⎪ LLh − 1/3 × KHh plength = ninth
⎪ LLh plength = long 3.3.2. Modeling the relations between human stature parameter and other
⎩ (10)
dimensions in height direction
where, WHh = a1 ∗ ST + b1 (11)

JL is jeans' length; HHh = a2 ∗ ST + b2 (12)

DWHh is human body's drop of waist-to-hip; LLh = a3 ∗ ST + b3 (13)
pwaist is the parameter of waist design; it includes three values: high,
normal and low waist (refer to Fig. 3); KHh = a4 ∗ ST + b4 (14)
plength is the parameter of jeans' length design; it includes six values: where,
hot, third, fifth, seventh, ninth and long (refer to Fig. 6);
LLh is human body leg length; ST is human body stature;
KHh is human body knee height; WHh is human body waist height;
d is a positive constant given by the designer. HHh is human body hip height;
LLh is human body leg length;
Definitions (8), (9) and (10) are used for the parametric design of KHh is human body Knee height;
jeans’ length.

216
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 7. Relations between jean pattern's dimensions and human anthropometric measurements in width direction. (Note: Wj is jean Waist Girth; Wh is human body
Waist Girth. Hj is jean Hip Girth; Hh is human body Hip Girth. e w is ease allowance at waist; eh is ease allowance at hip; ek is ease allowance at knee.).

a1 is regression coefficient between stature and human body waist Wh is human body Waist girth;
height; Hj is jeans' Hip girth;
a2 is regression coefficient between stature and human body hip Wh is human body Waist girth;
height; e w is Ease allowance at waist; its value depends on the design re-
a3 is regression coefficient between stature and human body leg quirement;
length; eh is Ease allowance at hip; its value depends on the design re-
a4 is regression coefficient between stature and human body knee quirement.
height;
b1 is constant term of regression equation the between stature and Human body hip girths are highly correlated to other girth direction
human body waist height; dimensions (Liu et al., 2017b; Liu et al., 2016b). Thus, the girth di-
b2 is constant term of regression equation the between stature and rection dimensions can be represented by hip girth using linear re-
human body hip height; gression.
b3 is constant term of regression equation the between stature and
Kj = aHj ⇒ Kj = a (Hh + eh) (17)
human body leg length;
b4 is constant term of regression equation the between stature and where,
human body knee height;
Kj is jeans' Knee girth;
Equations (11)–(14) construct the linear regression relations be- Hj is jeans' Hip girth;
tween stature and waist height, hip height, leg length, and knee height. Hh is human body Hip girth;
Other dimensions related to jeans' pattern-making in height direction eh is Ease allowance at hip; its value depends on the design re-
can also be represented by stature. For example, the drop of waist-to- quirement.
hip is equal to waist height minus hip height; crotch depth is equal to
waist height minus leg length, etc. Finally, all dimensions of jeans’ Finally, all dimensions of jeans’ patterns in girth direction are re-
pattern in height direction can be represented by stature parameter presented by hip girth using unary linear regression equation like
through these regression equations like Equations (11)–(14). Equation (17).

3.3.3. Modeling the relations between human hip parameter and other 3.4. Modeling the relations between adjustable parameters and jeans’
dimensions in girth direction pattern dimensions
According to the analysis in Fig. 7 (b), the relations between human
body and jeans are as follows: The adjustable parameters are used for adjusting jeans' patterns
Wj = Wh + e w (15) with which are the consumer is dissatisfied following a virtual or real
try-on. For example, if a consumer feels the jean is too long through try-
Hj = Hh + eh (16) on, we can adjust the length parameters to shorten the jeans' length. In
our previous analysis, jeans’ length, waist height type, leg opening girth
where,
and knee girth are selected as the adjustable parameters to construct the
proposed system. Thus, we classified dissatisfaction for each part into
Wj is jeans' Waist girth;
six degrees on a scale and pattern modification rules are shown in

217
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Table 1
Pattern modification rules.
Adjustable type Modification Parts IF THEN

Appearance adjustment Length Very long Decreasing 3ω% in length

Long Decreasing 2ω% in length
Little long Decreasing ω% in length
Little short Increasing ω% in length
Short Increasing 2ω% in length
Very short Increasing 3ω% in length
Waist height Very high Decreasing 3ω% in waist height
High Decreasing 2ω% in waist height
Little high Decreasing ω% in waist height
Little low Increasing ω% in waist height
Low Increasing 2ω% in waist height
Very low Increasing 3ω% in waist height
Knee Very wide Decreasing 3ω% in knee
Wide Decreasing 2ω% in knee
Little wide Decreasing ω% in knee
Little narrow Increasing ω% in knee
Narrow Increasing 2ω% in knee
Very narrow Increasing 3ω% in knee
Leg opening Very big Decreasing 3ω% in leg opening
Big Decreasing 2ω% in leg opening
Little big Decreasing ω% in leg opening
Little small Increasing ω% in leg opening
Small Increasing 2ω% in leg opening
Very small Increasing 3ω% in leg opening
Comfort adjustment Waist Very tight Decreasing 3ω% in waist girth
Tight Decreasing 2ω% in waist girth
Little tight Decreasing ω% in waist girth
Little loose Increasing ω% in waist girth
Loose Increasing 2ω% in waist girth
Very loose Increasing 3ω% in waist girth
Hip Very tight Decreasing 3ω% in hip girth
Tight Decreasing 2ω% in hip girth
Little tight Decreasing ω% in hip girth
Little loose Increasing ω% in hip girth
Loose Increasing 2ω% in hip girth
Very loose Increasing 3ω% in hip girth
Crotch Very tight Decreasing 3ω% in crotch width
Tight Decreasing 2ω% in crotch width
Little tight Decreasing ω% in crotch width
Little loose Increasing ω% in crotch width
Loose Increasing 2ω% in crotch width
Very loose Increasing 3ω% in crotch width

Note: ω is an adjustable parameter. It can be adjusted according to the customer's perception of the length and tightness of the garment. Its value defaults to 10 in
our software.

Table 1. system can generate jeans' patterns automatically. If the jeans' pattern is
not satisfying after virtual or real try-on, the adjustable input para-
meters can be applied to adjust the patterns’.
4. Application and validation of jeans’ patterns parametric design Using the proposed system, various jeans' patterns can be generated
at the conditions of different silhouettes, waist heights, jean lengths,
4.1. Application of jeans’ pattern parametric design human body waist girths, human body hip girths, statures, as shown in
Fig. 9. The results indicate that our proposed system can generate jeans’
In order to realize the proposed system, a data collection of an- patterns according to different body dimensions and styles.
thropometric measurement, which were involved in 120 young women
in northeast area of China, was conducted to calculate regression
coefficients a1, a2 , a3, a4 and constant terms b1, b2 , b3, b4 in 4.2. Validation of jeans’ pattern parametric design
Equations (11)–(14). An application named jeans' pattern re-
commendation system 2016 was developed according to formula (1), The aim of this research was to develop garment patterns rapidly;
(2), …, (17) and modification rules in Table 1 using visual basic com- therefore, it was necessary to test the patterns generated by the pro-
puter language (Fig. 8). In the proposed system, the default values of posed system for fit. In order to validate the proposed method, we
a, α, β, γ , eh, ew , ω, a1, a2 , a3, a4 , b1, b2 , b3 and b4 are shown measured and recorded a person's stature, hip girth and waist girth;
in Table 2. If fashion designers are not satisfied with the patterns by 165 cm, 68 cm and 88 cm respectively. Then we input these dimensions
these defaults, they can adjust them until they are satisfied. As shown in into the developed system in Fig. 8 and selected jeans' length as long,
Fig. 8, three geometric constraint parameters (jeans silhouette type, silhouette as straight, and waist height as normal. The final patterns
length and waist height) and three-dimensional constraint parameters were generated as shown in Fig. 10 (a). Using the virtual and real try-
(human body stature, waist girth and hip girth) are designed as the on, the jean was tested by a designer and the person. We used virtual
initial input items; and four adjustable parameters (jeans' length, waist try-on to check whether jean style meets design requirements. The jean
height, knee and leg opening) are designed as the adjustable para- fit is tested by real try-on. If virtual try-on shows that the jean style
meters. By inputting human body dimensions and a selected style, this meets design requirements, we will carry out real try-on to evaluate

218
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 8. Interactive interface of jeans pattern recommendation system 2016.

Table 2 significantly. Even when the garment patterns generated by our pro-
Constants' default values in this paper's formulas. posed method, do not completely meet pattern makers’ requirements,
Item a α β γ eh ew ω a1
they can adjust it easily. For example, if a pattern maker is satisfied
Default value 0.4 4 6 6 15 2 5 0.75 with the whole pattern generated by our method, except the part of the
Item a2 a3 a4 b1 b2 b3 b4 pocket, the only part to be modified is the pocket. As a result, the
Default value 0.69 0.53 0.37 −19.34 −28.21 −14.48 −15.91 materials and time are saved by not requiring to completely make a new
one.
Note: the unit of α, β, γ , eh , e w , b1, b2 , b3 and b4 is centimeter, and ω , a1,
a2 , a3 and a4 are dimensionless.
5. Conclusion
garment fit. If virtual try-on shows that the fashion style doesn't meets
design requirements, we will regenerate patterns until the fashion style In this paper, we proposed a parametric design method of garment
meets design requirements. Results show that the patterns meet the pattern based on body dimensions. Based on this method, we developed
subject's and design requirements. The time of pattern generation took a recommendation system of jeans' pattern. The jeans' patterns are
only several seconds. Through combing the traditional pattern-making parameterized by three geometric constraint parameters (jeans silhou-
method, our method can improve the efficiency significantly. ette type, length and waist height type) and three-dimensional con-
In the actual design process, a pattern maker makes patterns for a straint parameters (human body stature, waist girth and hip girth).
complex style in at least one day; however, using our method, only a Through combing the traditional manual pattern-making method, the
few seconds are required. The pattern-making efficiency improves parametric design method can generate jeans' patterns rapidly and

219
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 9. Examples of different styles and dimensions' patterns are generated by our proposed system.

automatically. Our proposed method saves plenty of time for pattern to parameterize its patterns. How to parameterize complex styles base
makers, and the jean's development cost reduces distinctly. Even per- on body dimensions will be a research direction. Moreover, we only
sons with limited or no pattern-making knowledge can also develop used linear models to construct relations between body dimensions and
professional garment patterns using our proposed system. garment patterns in this study. The relations between body dimensions
Although our research can generate garment patterns automatically and garment patterns are not entirely linear. In the future, non-linear
with parametric design method without expert knowledge, it still has models, such as neural networks, can be used to construct the mathe-
some limitations. For a complex fashion style, it will take a lot of work matical relations between them.

220
K. Liu, et al. International Journal of Industrial Ergonomics 72 (2019) 212–221

Fig. 10. Jeans pattern and try-on.

Conflict of interest Int. J. Cloth. Sci. Technol. 24 (5), 350–362.

Li, J., Lu, G., Liu, Z., Liu, J., Wang, X., 2013. Feature curve-net-based three-dimensional
garment customization. Textil. Res. J. 83 (5), 519–531.
We declare that we do not have any commercial or associative in- Liu, K., Kamalha, E., Wang, J., Agrawal, T.-K., 2016a. Optimization design of cycling
terest that represents a conflict of interest in connection with the work clothes' patterns based on digital clothing pressures. Fibers Polym. 17 (9),
submitted. 1522–1529.
Liu, K., Wang, J., Hong, Y., 2017a. Wearing comfort analysis from aspect of numerical
garment pressure using 3D virtual-reality and data mining technology. Int. J. Cloth.
Acknowledgments Sci. Technol. 29 (2), 166–179.
Liu, K., Wang, J., Kamalha, E., Li, V., Zeng, X., 2017b. Construction of a body dimensions'
prediction model for garment pattern making based on anthropometric data learning.
This paper was financially supported by the National Natural
J. Text. Inst. 108 (12), 2107–2114.
Science Foundation of China, China (No. 61806161), Special Scientific Liu, K., Wang, J., Zeng, X., Tao, X., Bruniaux, P., Edwin, K., 2016b. Fuzzy classification of
Research Plan Projects of Shaanxi Education Department, China (No. young women's lower body based on anthropometric measurement. Int. J. Ind. Ergon.
55 (5), 60–68.
18JK0352), Social Science Fund Project of Shaanxi Province, China
Liu, K., Wang, J., Zhu, C., Hong, Y., 2016c. Development of upper cycling clothes using
(NO. 2018K32), China National Endowment for the Arts, China (No. 3D-to-2D flattening technology and evaluation of dynamic wear comfort from the
2018-A-05-(263)-0928) and The Youth Innovation Team of Shaanxi aspect of clothing pressure. Int. J. Cloth. Sci. Technol. 28 (6), 736–749.
Universities, China. Liu, K., Zeng, X., Bruniaux, P., Tao, X., Yao, X., Li, V., Wang, J., 2018a. 3D Interactive
garment pattern-making technology. Comput. Aided Des. 104, 113–124.
Liu, K., Zeng, X., Bruniaux, P., Wang, J., Kamalha, E., Tao, X., 2017c. Fit evaluation of
Appendix A. Supplementary data virtual garment try-on by learning from digital pressure data. Knowl. Based Syst. 133,
174–182.
Liu, K., Zeng, X., Tao, X., Bruniaux, P., 2019. Associate design of fashion sketch and
Supplementary data to this article can be found online at https:// pattern. IEEE Access 7, 48830–48837.
doi.org/10.1016/j.ergon.2019.05.012. Liu, K., Zeng, X., Wang, J., Tao, X., Xu, J., Jiang, X., ... Bruniaux, P., 2018b. Parametric
design of garment flat based on body dimension. Int. J. Ind. Ergon. 65, 46–59.
Satam, D., Liu, Y., Lee, H.J., 2011. Intelligent design systems for apparel mass customi-
References zation. J. Text. Inst. 102 (4), 353–365.
Xiu, Y., Wan, Z.-K., Cao, W., 2011. A constructive approach toward a parametric pattern-
Cho, Y., Komatsu, T., Inui, S., Takatera, M., Shimizu, Y., Park, H., 2006. Individual pat- making model. Textil. Res. J. 81 (10), 979–991.
tern making using computerized draping method for clothing. Textil. Res. J. 76 (8), Yang, Y., Zhang, W., Shan, C., 2007. Investigating the development of digital patterns for
646–654. customized apparel. Int. J. Cloth. Sci. Technol. 19 (3/4), 167–177.
Kamalha, E., Zeng, Y., Mwasiagi, J.I., Kyatuheire, S., 2013. The comfort dimension; a Zhang, Y., Luximon, A., Pattanayak, A.K., Zhang, M., 2012. Shoe-last design exploration
review of perception in clothing. J. Sens. Stud. 28 (6), 423–444. and customization. J. Text. Inst. 103 (5), 541–548.
Kim, S., 2012. Simulation of bespoke garments using parametrically designed patterns.

221