Personalization of a Human Body Model Using Subject-Specific Clothing Industry Dimensions

Virtual human body models contribute to designing safe and user-friendly products through virtual prototyping. Anthropometric biomechanical models address different physiques using average dimensions. In designing personal protective equipment, biomechanical models with the correct geometry and shape shall play a role. The presented study shows the variations of subject-specific anthropometric dimensions from the average for the different population groups in the Czech Republic and China as a background for the need for personalized human body models. The study measures a set of clothing industry dimensions of Czech children, Czech teens, Czech adults and Chinese adults and compares them to the corresponding age average, which is represented by a scaled anthropometric human body model. The cumulative variation of clothing industry dimensions increases the farer is the population group from the average. It is smallest for the Czech adults 7.54% ± 6.63%, Czech teens report 7.93% ± 6.25% and Czech children differ 9.52% ± 6.08%. Chinese adults report 10.86% ± 11.11%. As the variations of the particular clothing industry dimensions from the average prove the necessity of having personalized subject-specific models, the personalization of particular body segments using the measured clothing industry dimensions leading to a subject-specific virtual model is addressed. The developed personalization algorithm results in the continuous body surface desired for contact applications for assessing body behavior and injury risk under impact loading.


26
Virtual prototyping enables fast product design and optimization using numerical 27 calculation. Virtual human body models play an important role in the design of human-28 friendly and safe personal products. Addressing the wide population, customized 29 human body models [1,2] have a wide range of applications. A virtual human body 30 model with biofidelity can also serve for assessing injuries caused by external impacts [3]. 31 The personalization of such model is usually based on mesh morphing [12], which 48 needs to have a considerable number of corresponding landmarks defined on both the 49 reference and the target models [13,14]. 50 Both approaches are based on a reference model, which is a necessary starting point 51 for the development of a subject-specific model. The scaling approach is simpler as 52 it addresses only global (averaged) anthropometric dimensions for target population 53 groups [7]. Scaling is a perfect tool for creating generic human body models representing 54 a population group for virtual prototyping [8] as they have the correct mass distribution 55 and body flexibility to cover a wide spectrum of the population. However, if we want to 56 address the designing of PPE for a particular person, a subject-specific (personalized) 57 approach is needed as scaling does not provide details of the particular subject even 58 inside a particular population group. Additionally, the anthropometry of particular 59 segments might be unrealistic for a different population groups when using only scaling.

63
The purpose of the paper is to measure subject-specific anthropometric dimensions 64 coming from the clothing industry for designing protective garments [15] for different 65 population groups. The subject-specific dimensions are compared to the average an-66 thropometry, which is obtained by scaling the reference anthropometry [16] using the 67 average anthropometric data for the population group [17]. Based on the measured 68 dimensions, a personalization algorithm for creating a subject specific human body 69 model using only the clothing industry dimensions is developed. 70 There are also other initiatives [18] and options (e.g. 3D scanning) to generate a 71 subject-specific model using three-dimensional scanning, however, the paper benefits  were carried out further. Their total number was 68 ranging in age from 24 to 60 years. 84 The males and females were equally represented. 85 The measurements were carried out on adult students and adult employees (N =  The clothing industry anthropometric dimensions [20,21] as shown in Figure 1 were 104 measured using a tailor measuring tape on the particular body segments.
105 Figure 1. Clothing industry anthropometric dimensions. Identified clothing industry dimensions are adopted from the protective garment clothing industry [15].
The anthropometric dimensions are described in Table 2. The dimensions are 106 mentioned by their number in the parenthesis in the further text. The age and the total 107 weight of each subject were also reported.
108 Table 2. Clothing anthropometric dimensions. Lengths between particular external body points and the circumference of particular segments. the previously developed and published algorithm [7,22].

153
The variation of the subject-specific dimensions from the average anthropometry  The reference model is a previously developed and published virtual biomechanical 160 human body model [22] already implemented in a computational environment for 161 assessing the risk of injury during an impact. Figure 2  The personalization method developed in the paper upgrades the existing scaling The personalization provides the local geometrical and biomechanical details of the 186 human body. Firstly, the scaling [7] is done for each subject to target the given gender, 187 age, height and weight. Having the scaled model, the personalization to provide correct 188 anthropometric dimensions is carried out. The two-step approach is described as (1) directs laterally from the right to the left forming the right-hand side system.

197
Personalization is the next step after scaling. Scaling is a necessary step to have 198 a model with global dimensions for the particular human body size before applying 199 the mathematical operation to refine the surface locally. As the scaling algorithm was 200 published previously, this paper will not address it, referring to the work by [7]. For per-201 sonalization purposes, the scaled model is divided into segments summarized in Table 2 202 and schematically shown in Figure 3. Each segment can be described by the anthropo-203 metric dimensions from Table 2 displayed in Figure 1. The personalization algorithm 204 runs independently of each of the 16 subject-specific segments defined in Table 3. using the corresponding dimensions from Table 2. There is a reference node X s S0 chosen 206 on each segment as a local coordinate system origin, against which it is personalized, and Any node X s Si , i ∈ {1, . . . , m} (m being the numbers of nodes on the segment s) on 218 the scaled model S is personalized as

205
where X s Ti is the personalized node on target model T and R s = R s (X) is the transforma- where p s i for i ∈ {x, y, z} are the personalization coefficients at location X in the global 227 coordinate system for segment s. We use polynomials of degree n s i in the form to interpolate the personalization coefficient along each axis i ∈ {x, y, z}. Lower index j r 229 for r ∈ {1, . . . , n s + 1} is chosen from set {1, . . . , 23} as numbered according to Table 2. anthropometric dimensions (see Figure 1) as where lower indexes T and S mean target (personalized) and scaled dimensions, respec-235 tively, and j ∈ {1, . . . , 23} is the dimension number as numbered according to Table 2 236 and shown in Figure 1. For simplification in the further text, we indicate the personal- polynomial P(17-16) indicates a constant polynomial calculated as the ratio polynomial P (13,14) indicates a linear polynomial interpolating the personalization 243 coefficient between points and polynomial P(6,7,3,4,5) indicates a polynomial of degree 4 interpolating the person-245 alization coefficient between points The polynomial interpolation might suffer from inaccuracy and unrealistic over-247 shoots in the polynomial shape if we interpolate incomparable values, which is why the 248 scaling is carried out in the first step to approach the target shape, which is refined by 249 personalization.

250
For each segment, we firstly personalize the height based on the vertical dimensions

255
The frontal (segment depth) and lateral (segment width) personalization depends on 256 the circumference dimension of the particular segment. The circumference dimensions 257 are (2), (3), (4), (5), (12), (13), (14), (15), (18), (19), (20), (21) and (23) as shown in Figure   258 1 by means of numbering according to Table 2. As many segments are described by 259 circumference at the different horizontal location, the polynomials of the degree n > 0 260 interpolate the personalization coefficient along axes x and y depending on axis z, which 261 leads to the exact circumferences according to Figure 1 in the personalized model, so where i ∈ {x, y} and n s is the polynomial degree. We use the same polynomial degree n s 263 for axes x and y for segment s, because the polynomials are formed by the same number 264 of circumference dimensions for each segment.
265 Figure 1 also defines additional width dimensions, particularly (6) and (10), which 266 are useful as additional measures for personalization. The personalization of each 267 segment is described in detail in the following paragraphs.  The neck follows the thoracic at the joint between vertebrae C7 and T1 [20], which is 293 also the reference point. According to Figure 1, the height is personalized by a constant 294 coefficient calculated as the ratio of the total body height without shoes (1), from which 295 we subtract the back length from neck to waist (9) and the abdominal height (22). As  The head follows the neck at the clivus [20]. As height 1-9-22 is the total height of 305 the neck and the head, the same personalization coefficient is used for the head height.

306
The height personalization uses the same personalization coefficient as the neck uses.

307
The head size is not typically measured for the means of the clothing industry, so the 308 coefficients are missing. Therefore, the depth and width are personalized by a constant 309 coefficient using the head circumference (23), which is measured.   The thigh follows the abdomen in the hip joint [20]. According to Figure 1, the height 330 is personalized by a constant coefficient addressing the ratio of the length from crotch 331 to knee (16). The depth and width are personalized using the same cubic polynomials 332 interpolating the ratios between the hip circumference at the widest point (5) The calf follows the thigh in the knee joint [20]. The height is personalized by a  The foot follows the calf in the hip ankle joint [20]. As the foot is not typically 342 measured in the clothing industry and its small size, mass and inertial effects do not 343 considerably influence any dynamical action, the foot is not personalized and only the 344 dimensions raising from of the scaling are taken into account.

345
As the human body is expected to be symmetric, the personalization of left and 346 right arms, forearms, palms, thighs, calves and feet is the same.  As the reference human body model is developed for dynamical analysis, the last 350 step is updating the masses and inertias of the particular segments concerning the shape 351 change as well as the ranges of the particular joints concerning the age using the same 352 approach as described and published by [7].

353
The personalization process is implemented in Python. The Python script first 354 reads the structure of the reference model as the finite element model [22]. The structure 355 concerns the multi-body system, which separated the particular segments and their 356 nodes representing the geometry into rigid bodies. The next step is the scaling of nodal 357 coordinates and updating the stiffness of the joints, which is is updated independently 358 using the approach previously developed and published [7,25].

359
After the scaling of each segment, the personalization runs. Then, the segment 360 masses and inertia are re-calculated based on the particular segment volume and shape 361 change using the approaches previously developed and published [7]. After that, the  2. Scale reference model.

375
The optimization loop is illustrated in Appendix A as a flowchart.   All the measured subjects were reconstructed by scaling using the previously devel-385 oped and published scaling algorithm [7] and the reference model [22]. The following  Figure 4, the group of 32 Czech teens in Figure 5, the group of 50 Czech 393 children in Figure 6 and the group of 68 Chinese adults in Figure 7.   The displayed models are set to scale so that they can be compared with each other.

411
By nature of the personalization algorithm targeting the particular dimensions from 412 Table 2, the developed subject-specific virtual human body models fit to their real subject 413 by those dimensions.

415
The presented study shows the variations of the subject-specific anthropometric  The variation of the Czech adults from the average (Figure 4) is the lowest in all 421 groups and is around 7.54 % (8.2 % for males and 6.9 % for females). This is probably 422 caused by the fact that the reference model is an average European subject, which is 423 close to the adult anthropometric group. The greatest variation is in the shoulder width 424 (6) followed by the wrist circumference (14) and the frontal shoulder height (8). The 425 great variation in the thoracic circumferences and the frontal thoracic height is probably 426 caused due to the fact that the real shape distribution can be affected by the muscle 427 mass or the breast size of the particular subject. The lowest variation is in the knee 428 circumference (19) followed by the arm circumference (12) and the hip circumference (5) 429 meaning that those dimensions can be well described by average anthropometry.

430
Decreasing the age, there is a slightly greater variation equal to 7.93 % (8.15 % for 431 males and 7.9 % for females) for the Czech teens ( Figure 5). The teens, as well as the 432 adults, are well described by the average European subject. The greatest variation here is 433 again in the shoulder width (6), followed by the wrist circumference (14) and the frontal 434 thoracic height (8). Again, all these measures can be affected by muscle mass. The lowest 435 variation is in the knee circumference (19) followed by the waist circumference (4) and 436 the abdominal circumference (3).

437
The average variation for the Czech children ( Figure 6) equals 9.52 % (9.57 % 438 for males and 9.49 % for females). Here the variation is greater as the children have 439 different anthropometry and segment proportions compared to the adults. However, 440 the variation is not so great with the biggest in the wrist circumference (14) followed by 441 the shoulder width (6) and the ankle circumference (21). The lowest variation is in the 442 head circumference (23) followed by the shoulder length (10) and the thigh length (16).

443
The greatest average variation is seen in the group of Chinese adults ( Figure 7)

466
The current paper provides a measurement of 23 major clothing industry dimen-467 sions on 220 human subjects including children, teens and adults in the Czech Republic 468 and China. The number of Czech subjects was 152, ranging in age from 6 to 56 years.

469
Additional measurements on 68 Chinese adults ranging in age from 24 to 60 years were 470 carried out. The males and females were equally represented.

471
Using a single reference model, the subject-specific human body model is developed.

472
On each subject, the scaling procedure was carried out to develop the generic scaled 473 model related to the particular gender, age, height and weight. The anthropometric 474 dimensions adopted from the clothing industry were compared to the measured ones.