Academic Journal of Science and Technology 
ISSN: 2771-3032 | Vol. 4, No. 3, 2022 

 

73 

Research on Young Females' Body Classification Based 
on SVM 
Suyu Zhang1, * 

1 College of Artificial Intelligence, Wenzhou Polytechnic, Wenzhou 325035, China 
* Corresponding author: SuYu Zhang (Email: 23209695@qq.com) 

 
Abstract: Young females’s data in Jiangsu and Zhejiang areas are acquired by using 3D Body Scanner made by American 
TC2. Then according to the four parts of body shape character, we can classify human data into different kinds of women’s body 
types. In the process, SVM(support vector machine) model is used to identify the young women’s shape. The results show that 
SVM is an effective identified and classified method due to short time and high precision. The reference data can be provided 
for the garment pattern design of young females’s apparel. 

Keywords: SVM, Young female, Human classification. 
 

1. Introduction 
Nowadays, clothing fit has been widely regarded as an 

important indicator that affects the appearance and comfort of 
clothing. According to a survey by KSA (Kurt Salmon 
Associates): 50% of women complain about not being able to 
buy fit clothes. Another survey indicated that 50% of the 
reasons for the return of goods were caused by inappropriate 
clothing. [1] Body shape directly affects structural design. 
Therefore, the purpose of body classification is to provide 
basis for clothing style design and structural design. 
Therefore, the accuracy and timeliness of body classification 
algorithms are crucial. 

Body type analysis plays an important role. It is not only 
one of the key technologies to realize MTM(Made to 
Measure), but also an important procedure to combine 3D 
body scanner with digital garment production. Body type 
analysis means to analyze the data in Anthropometrical 
Database, then acquire parameters by computing the data 
based on the statistical theories, finally classify human body 
characteristics using clustering analysis. It provides 
advantageous help to the mapping from body form to clothing 
size and other application of MTM [2]. 

But the description of body character involves the 
relationship between data and specific judgment, which is 
hard to depict by traditional mathematic methods because of 
the non-linear relation among parameters and the existence of 
multi influencing factors. So we adopt Support vector 
machine (SVM) which is the very method to resolve this kind 
of problems. By checking the model through actual body 
measurements, the method in this paper can fully satisfy the 
need of body type analysis in MTM. 

SVM is a learning method proposed by Vapnik et al. based 
on statistical learning theory. In recent years, there have been 
many applications in pattern recognition, regression analysis 
and feature extraction [2]. According to Vapnik's principle of 
structural risk minimization, SVM selects a limited data set to 
provide the optimal model with the minimum expected risk. 
Moreover, SVM algorithm is a convex quadratic optimization 
problem, which can guarantee that the extreme value solution 
found is the global optimal solution. Some recent applications 
show that the SVM method shows better adaptability and 
popularization ability than traditional methods including 

neural network methods [3,4]. 
The rest of the paper organized as follows: Section 2 

describes the general structure of the SVM model; Section 3 
explains the data source and classification; finally some 
concluding remarks are drawn from section 4. 

2. Theory and Arithmetic of SVM 
SVM take the following advantages over conventional 

statistical learning algorithms: (1) High generalization 
performance even with high dimension feature vectors; (2) 
The ability to manage kernel functions that map input data to 
higher dimensional space without increasing computational 
complexity. 

2.1. Optimal Hyperplane  
The main idea of SVM is to create an optimal hyperplane 

to classify the data into two classes (positive and negative) 
and maximize distance between the hyperplane separating the 
two classes and the closest data points to the 
hyperplane.(Figure 1)[4] 

 

 
Figure 1. Optimal classifying line under linear classified 

condition 
 

Figure 1. shows training examples linearly separated into 
two classes. 

It starts with a set of l training data (x1,y1), ,(xl.yl), where 
n

ix R is an n-dimensiona vector and { 1, 1}iy is the 
class label of its data: 

f(x)=-1 

f(x)=0 

Optimal hyperplane 
 hyperplane 

Margin=2/||w|| 

Origin 

f(x)=1 



 

74 

 1 1( , ),x y Rn R 
The basic idea of SVM is to maximize the margin between 

the positive and negative examples. Firstly, the training 
examples are mapped from original spaces Rn to feature 
spaces 1 2( ) ( ( ), ( ), ( ))x x x x   by a non-linear 
function (  ). Secondly, non-linear estimated function is 
changed into linear estimated function in hyperplane feature 
spaces by the most optimal decision function, as follows: 

 ( ) ( )y x w x b  
Finding w and b is just minimizing R: 
 

2 *

1

1min ( , ) || || ( )
2

R w w c        (1) 

 
s.t. i i i( )y w x b  

 * *( ) , , 0, 1,i i i i iw x b y i l  
in Eq(1): ||w|| Controlling the complication of models 
        C  Const 
       Misclassification errors 
The answer of Eq(1) can be got by Lagrang Function: 
 

  

* * * *

1

*

1 1

* * *

1

1( , , , , , , , ) ( ) -
2

(( ( )) - ) - ( -

( ( )) - ) - ( )

l

i i
i

l l

i i i i i i
i i

l

i i i i i i
i

L w c w w c

a w j x y b y

w j x b

 (2) 

 
According to optimal conditions: 

*0, 0, 0, 0
i i

L L L L
w b

 

The results are : 
 

* *

1 1

* *

( ) ( ), ( ) 0,

0, 0, 1,

l l

i i i i i
i i

i i i i

w x

c c i l
       (3) 

Defining the Kernel Function based on Mercer’s conditions: 
 

 ( , ) ( ) ( )i j i jk x x x x              (4) 
 
Through Eq(2),Eq(3) and Eq(4), the optimization problem 

is exchanged as follows: 
 

 
* *

, 1

* *

1 1

1max *)=- ( )( ) ( , )
2

( ) ( )

l

i i j j i j
i j

l l

i i i i i
i i

W K x x

y

   (5) 

 

s.t.  
*

*

( ) 0

0 ,0 , 1,
i i

i ic c i
 

So the final non-linear estimated function is: 

   *( ) ( ) ( , )i j i jf x K x x b         (6) 

2.2. Kernel Function 
We choose different Kernel functions that can construct  

variable support vector products. In general, Kernel function 
including: 

1)Polynomial Kernel Function: ( , ) [( , ) 1]d
i j i jK x x x x ; 

2)Radial Machine Kernel Function:
2

2

|| ||

2( , )
i jx x

i jK x x e ;
3)Neural Network Kernel Function: 

( , ) ( ( , ) )i j i jK x x S v x x c  
In this paper, we use radial machine Kernel functions that 

have been very effective. 

2.3. SVM Network Structure 
The realization of SVM in this paper due to the structure as 

Figure 2 shows: 

 
 

Figure 2. SVM network structure diagram 
 

Where αi α*i is the weight of network, variable 
x1,x2,…,xl is introduced for input numbers and y is output 
number. 

Systematic identification based on SVM steps as follows: 
1)Choosing the training examples, modifying and pre-

computing the data; 
2)Using the function of SVM to train the chosen data: input a 

set of l training data x1,x2,…,xl, then output variable y;  
3)Constructing the modeling of SVM and predicting the 

unknown examples. 

3. Experiment 
3.1. Analysis on Body Characteristic Parts 

According to garment pattern design and the experience of 
fashion designers, there are more than 80 geometric 
characteristic sizes [6,7]. Among them the following four 
parts, flank shape, bust shape, hips shape and abdomen shape 
can reflect the most of the information about the body form, 
and are also the most important geometric characters of 
human body model [8,9,10]. 
  The flank part can reflect information which the front part 
and back part can not reflect. For example, vertical section 
body form curve can clearly reflect whether the body is 
prominent bust, bend back or not. Bust and back will 
influence the length of front centerline and back length, and 
then result in the changes of patterns. 

The breast is the main factor influencing bust shape. Its size, 
shape and height directly affect breast dart design of the front 
piece.  

The hip part is an important part in the research on the 
fitting of the lower part. The angle from back waist to back 
hips is a parameter to design back rise. It can reflect the raise 

  

x1 x2 x

y 

 



 

75 

shape of hips, the fuller the hips, the larger the angle, and the 
thinner the hips, the smaller the angle. 

The abdomen part influences the waist dart. It is shown that 
the abdomen shape plays a more important part in the body 
form. It is easy to accumulate fat for abdomen, which will 

cause the change of the front of the lower part [11,12]. 
Summing up, we get the parameters depicting women’s 

body form as Table 1 shows. And this demonstrates the multi-
dimension of human body. 

 
Table 1. Parameters and Description of the Parts 

Parts Parameters Description 

 
 

Flank 

back bend index, the angle from shoulder blade to back 
neck, the angle from shoulder blade to back waist, Under 

bust slope angle, Front neck to breast angle 

back bend(C1) 
 

medium(C2) 
 

bust prominent (C3) 
 

 
 

Bust 
 

under bust slope angle, front neck to breast angle, side neck 
to breast/ breast to waist, bust to bust /bust front, bust 

width/bust thickness, bust full 

flat(R1) 
 

medium (R2) 
 

full(R3) 

 
Hip 

The angle from back waist to back hips, hips full, 
hips thickness/hips length 

flat (T1) 
medium (T2) 

full (T3) 

 
Abdomen 

the angle of abdomen prominence, waist full, abdomen full, 
waist length 

 
medium(F1) 

prominent (F2) 
 

3.2. Body Type Identification Method Based on 
SVM Algorithm 

Because garments enterprises haven’t analyzed the target’s 
body form, they always produce “safe” garments, which 
results that customers can not buy garments fitting their body 
form. Fully analysis on body form enables people to acquire 
more information when they are tailoring and transfer 
sensibility recognition to logos evaluation. This rationalizes 
the formulation of data. Therefore, if we can simulate and 
reconstruct the fashion designers’ experience and arrange the 
proportion of every body form by computer, the designers’ 
burden will be lessened. As the SVM is able to approximate 
the complicated non-linear relation of human body and 
simulate designers’ non-linear experience. In this way, we can 
determine the system category to be developed and garment 
production using the results of body form analysis. 

In this research, four machines are constituted on the basis 
of flank shape, bust shape, hips shape and abdomen shape and 
their optimal structures are acquired. we take the 
classification of flank shape as an example. 

1) Ascertain the collection of training samples: In this 
experiment, we choose 350 young women age from 18 to 35, 
average height 160.55cm, average weight 50.5kg and average 
age 23. Equipment: TC2 3D body scanner made in American, 
the revised altimeter and weight scale and so on. First, the 350 
young women’s bodies are scanned by 3D body scanner. Then, 
10 experts score the parts. The data of the 350 individuals’ 
bodies are made as input and the corresponding description as 
output, so the training sample collection is got.   

2) Pre-computing the data: Based on the analysis on human 
body form character and fully considering the need of 
garment pattern design, we can use the parameters which in 
Table 1 as the input of the SVM, they are: back bend index, 
the angle from shoulder blade to back neck, the angle from 
shoulder blade to back waist, Under bust slope angle, Front 
neck to breast angle. 10 experts are asked to give semantic 
description of the body form and their judgments “back 

bend ”, “medium” and “bust prominent” (the “description” in 
Table 1) are regarded as output vectors. 

The data are normalized in [-1 ,1].The normalization 
formula is  

  

              
min

2 1
max min

i i
i

i i

x x
x

x x
             (7) 

 

ix —— the original data of index variables.  

max { ix }—— the maximum of the same index data. 

min { ix }——the minimum of the same index data.). 
 

3) Constructing the modeling of SVM: In the process of 
training, there are 5 parameters of input and 3 categories of 
output in every training sample, which are 1, 2, 3. Accordingly, 
their goal outputs are [+1,-1,-1], [-1,+1,-1], [-1,-1,+1]. And 
we make the category corresponding to the maximal output 
as the category for classifying. Finally, 300 of young 
women’s bodies are regarded as training samples and 50 as 
testing samples. 

4. Conclusions and Future Works 
4.1. Conclusions  

1) The research in this paper demonstrates the feasibility of 
SVM applied to body type classification and prediction. It can 
avoid the partial minimal point and improve the precisions of 
results. The models have been applied to garment producing 
and the result shows the models have a good effect on the 
garment producing.  

2) As long as we find out the factors influencing figure 
characteristic and enough samples to be learned by the 
network, we can make accurate and applied evaluation on 
body type, and optimize the index. This is significant to 
garment enterprises because it provides them with basis to 
arrange producing garment size reasonably. 



 

76 

3) This paper provides powerful data support to the 
mapping between body form and garment size and solves one 
of the key technologies to realize MTM. The body type 
analysis algorithm in this paper satisfies the need of body type 
analysis; it will promote the realization of MTM in our 
country. 

4.2. Future Works 
First, we can make software to realize the identify methods. 

Second, more samples are need to make the model have more 
extensive applications and to reduce errors. Third, we can 
combine SVM with BP, or study on the parameter setting 
further. 

Acknowledgment 
The paper is supported by Zhejiang Provincial Department 

of Education named Research and Development of 3D Body 
Costume CAD System Based on OpenGL(FG2022048). The 
authors would like to thank the 10 experts for their 
contributions to this project and the former researchers. 

References 
[1] Michael T.Fralix.(2021). From Mass Production To Mass 

Customization. Journal of Textiles and Apparel, Technology 
and Management, 2, Pp.1-13. 

[2] XIE Ji-jian, LIU Cheng-ping, Theory and Application of Fuzzy 
Mathematics [M], Second Version, Wuhan: Huazhong 
University of Science and Technology Publishing House, 2020, 
Page No: 10-25 

[3] Karla Simmons, Cynthia.L. (2014). Female Figure 
Identification Technique (FFIT) For Apparel. Part I: 
Describing Female Shapes. Journal of Textiles and Apparel, 
Technology and Management, 4, Pp.1-16. 

[4] Suykens J A K Nonlinear Modeling and Suppoo Vector 
Machines[c] Proc of the 38th IEEE Conference on 
Instrumentation andMeasurement Technology 2021 287—
294 

[5] Lu Jian-Chang, Niu Dong-Xiao, Jia Zheng-Yuan. A Study Of 
Short-Term Load Forecasting Based On Arima-Ann, 
Proceedings of The Third International Conference On 
Machine Learning And Cybernetics, Volume:5, Issue:7, 
August, 2014, Pp.3183-3187. 

[6] Park D C, Ei-Sharkawi M A, Marks R J Li Et AL. Electric Load 
Forecasting Using An Artificial Neural Network[J]. Ieee 
Transactions On Power Systems, Volume: 6, Issue: 2, 2011. 
Pp.442-449.  

[7] Self.(2020,March). The best workout for you body type. 
Self.22 (3), Pp.65.  

[8] Tsang, B., Chan,C.K.&Taylor,G.(2020). Kin anthropometry: 
Study of the physique of disciplined personnel. International 
Journal of Clothing Science and Technology, 12(2). 

[9] LePechoux,B.(2008,September).Standard sizing and fit testing 
applied to women’s hosiery. Unpublished Doctoral A1 
requirement paper, North Carolina State University, College of 
Textiles.  

[10] Tamburrino, N.(2012, April). Apparel sizing issues-Part1. 
Bobbin. 33(8), Pp.44-46. 

[11] ZHANG Zu-fang,DAI Wei. Study on Characteristic 
Parameters on Female Torso Shape. 2014 Shanghai 
International Fashion Culture Festival. Pp.180-181,189. 

[12] Feng-yuan Zou, Xiao-jun Ding etc. Application of Neural 
Network to Identification of Young Females' Body Type. IEEE 
International Conference on Systems, Man, and Cybernetics, 
2016.