A virtual e-bespoke men's shirt based on new body measurements and method of pattern drafting

Abstract

Customization is prevailing in the apparel industry with increasing requirements from consumers and the popularization of new technologies. This study aimed to establish the novel approach of applying existing and new body measurements to customize the pattern block of a men's shirt, to enrich the anthropometric database, and to develop the fit evaluation procedure. New body measurements were extracted from 156 scanned male mesh bodies in accordance with the morphological features and developing method of pattern block sketching. Owing to these new body measurements, the customized shirt with assured high-level fit can be obtained by generating original patterns as bespoke, on the one hand, and by transforming ready-to-wear patterns, on the other hand. The first way is e-bespoke tailoring that utilizes the developed schedule of body morphological features, improved shirt pattern of desired style (body fit, slim fit, regular fit, and comfort fit), and virtual try-on software CLO 3D. The proposed method of virtual e-bespoke design allows readily completing a well-fitted and balanced men's shirt, which will contribute to the efficiency of customization and quality of end-products for the apparel industry.

Keywords

3D scanning morphology e-bespoke balance body measurement fit men's shirt

Customization, also known as made-to-measure (MtM) and bespoke, is reviving as the foremost trend in the apparel industry.¹ For one thing, the productivity of customized apparel has been highly improved owing to the comprehensive development of industry 4.0 along with the advanced technologies such as 3D body scanning, virtual reality, and artificial intelligence.^2,3 For another, the consumers shift more attention to personality, comfort, and fit with the upgrade of consumption and aesthetics.^4,5 The market of men's customized apparel is showing great potential for high consumption growth especially by young men, the shirt being one of the most prevalent categories.⁶ Among all the indicators, apparel fit is regarded as the most crucial one in determining consumers' overall satisfaction.⁷

An apparel fit is the indicator to judge the two final results—design and tailoring.⁸ Well-fitted apparel should be comfortable and not impede the movement of the wearer.⁹ To determine the fit, five fundamental factors are adopted as subjective evaluation criteria related to three different areas:

One factor relates to textile fabrics: the grain direction which influences fabric draping;

One factor relates to the pattern block: the ease allowances which show the difference between an apparel and body measurements and produce an air gap between them in the ready “body-apparel” system;

Three factors relate to the “body-apparel” system: the contour lines which reflect the relationship between silhouette, apparel construction, and human body; the set which is as an indicator of smoothness, location of stress folds, or unnecessary creases; the balance which shows the symmetry of apparel around and over the body in front, profile, and back views.¹⁰

The apparel fit relies on the generated patterns since the pattern design is implemented based on the real body morphology.

Ready-to-wear (RtW) apparel is manufactured on the basis of patterns designed for typical bodies from the standard sizes which are derived from anthropometric databases.^11,12 Concretely, seven basic body measurements from the sizing system are required as usual to draft RtW patterns of a men's shirt: height (H), chest girth (CG), waist girth (WG), hip girth (HG), neck base girth (NG), shoulder width (SW), and arm length (AL).¹³ However, consumers have continuously experienced fit problems due to an incongruent relationship between the garment from the RtW system and the body morphology.¹⁴ The primary reason is that very few individuals have the same size and shape as the standard model due to different body contour, posture, and proportion, which all affect the fit.¹⁵

In order to eliminate an unconformity between the body and the garment, MtM apparel could be produced in accordance with the morphology of a real body by adapting the RtW patterns. In general, most contemporary MtM services of men's shirts offer several alternative personalized options, including textile fabrics, styles, and decorative details (collar, pocket, cuff, trim, monogram, etc.). The style of shirt, also referred to as fit that indicates the specific silhouette and shape of shirt around the torso, is usually sorted into body fit (or extra slim fit), slim fit, regular fit, and loose fit (or comfort fit).^16,17 Nevertheless, an adoption of body measurements varies considerably in the degree of customization and the complexity, which accordingly determines the level of the end-product apparel fit.

There are three primary levels of customizing dimensions of men's shirts. First, only the same sizes as the RtW system and several fundamental body measurements (height, weight, NG, AL) are customizable.^18–20 Second, more body measurements (CG, WG, HG, SW, wrist girth (WRG), etc.) are included.^21–23 Third, information from the consumer about their morphological attributes, for instance shoulder sloping, volume of abdomen, posture, and so on, is required.^24–26 With more morphological information involved, MtM services provide the apparel fit infinitely superior to those of RtW system.

However, all three approaches are based on traditional body measurements and landmarks such as shoulder point (SP), side neck point (SNP), and others, which are inadequate to represent the essential morphological features. A good apparel fit cannot be invariably guaranteed, especially for untypical bodies.

With the advances in computer and scanning technologies, many problems of customized patterns related to digital numeralization of morphological features can be solved now. The non-contact 3D body scanning technology is capable of rapidly obtaining hundreds of body measurements and generating a precise 3D mesh model as the real scanned body which can be directly applied in apparel design.²⁷ Moreover, with contemporary 3D garment computer-aided design (CAD), realistic virtual try-on visualization is accessible in real time for various avatars, pattern blocks, and textile fabrics. An evaluation of apparel fit can be conducted with the models generated in virtual reality instead of repetitive experiments with real samples. These technologies are opening new possibilities of improving the customized apparel fit.²⁸

Many recent researchers proposed new methods for customizing pattern blocks. First, the 2D patterns can be transformed from 3D meshed patterns that can be created by establishing the group of connecting feature points on the scanned body surface.^29,30 Through this 3D-to-2D flattening technique, the customized patterns can be directly obtained without body measurements. However, this method is advantageous for formfitting or simple-styled close-fitting apparel.

Second, new parametric pattern-making models were proposed to automatically generate patterns of different apparel styles by inputting particular body measurements, which were still derived from existing measurements and landmarks.^31,32 Moreover, some scholars have established the estimation models by regression equations or artificial intelligence, which can instantly output the pattern-related body measurements (girths, heights, and widths) by inputting a few basic measurements, which were of high inaccuracy to estimate the measurements of untypical bodies.^33,34

Both models cannot entirely characterize the individual morphology, resulting in misfit of apparel. Although some scholars proposed the method for pattern adjustment to eliminate local misfit such as in the shoulder area, there is no systematic adjustive method for improving the holistic apparel fit.³⁵

According to the contemporary researches and services of MtM, 16 existing body measurements (BM_E) are commonly adopted in men's shirt customization to acquire the holistic apparel fit: CG, WG, HG, NG, SW or shoulder length (SL), AL, WRG, upper arm girth (UAG), surface length from SNP to front waistline (SNW_F), surface length from SNP to back waistline (SNW_B), bust height (H_B), waist height (H_W), hip height (H_H), cervical height (H_BNP), across front width (AW_F), and across back width (AW_B).^{24–26,36,37} Although the pattern can be further customized with BM_E, some morphological proportions and features are neglected. For example, CG, WG, and HG, which depict the primary figure, are incapable of reflecting the anteroposterior proportions.

Therefore, there is a high necessity of discovering the list of new body measurements (BM_N) which improve on characterizing the holistic individual body shape and the corresponding pattern customizing method based on BM_N for assured good apparel fit.

As the process of pattern design and fit evaluation are shifting from reality to virtual reality, conventional MtM is transformed into e-bespoke as the high-level digital customization.

To accomplish e-bespoke, this study aimed to propose new complementary body measurements, and to develop the methods of applying the measurements to men's shirt pattern drafting and numerical criteria of fit evaluation. With the 3D body scanner and 3D modeling software, the BM_E and BM_N of four primary body segments (torso, neck, shoulder, and arm) were extracted from the 156 scanned body mesh models. With these measurements, the shirt pattern is precisely drafted in accordance with body shape and desired shirt style. Meanwhile, 3D CAD was used to objectively and subjectively evaluate the realistic try-on visuals of the generated patterns without repetitive real wearing trials. The final customized virtual shirts specializing in different bodies, styles, and textile materials exhibited assured high-level apparel fit as a result of compatibility of pattern and body.

Methods

Subjects

A total of 156 young men (94 male students from central China and 62 male students from west Russia) aged from 18 to 30 years without morphological malformation (such as bone, joint, and muscle deformity) and physical disability were voluntarily enrolled for body scanning as the pilot research. The scanning process was conducted by VITUS Smart XXL non-contact 3D body scanner (Human Solutions, Germany) at Wuhan Textile University and Ivanovo State Polytechnic University. In accordance with ISO 20685-2010(E),³⁸ every volunteer was scanned for several times until the scanned model was applicable for subsequent measuring and try-on processes. The selected subjects were 95 men of Y-type (drop value is from 17 to 22 cm), 55 men of A-type (drop value is from 12 to 16 cm), and 6 men of B-type (drop value is from 7 to 11 cm) according to the Chinese sizing system.

For the sake of the validity of subsequent experiments, the least sample size was calculated as equation (1)³⁹

n = \frac{Z^{2} \times S D^{2}}{d^{2}}

(1)

where n is the estimated sample size, Z is the standard normal variate (in this study Z = 1.96 for 95% confidence interval), SD is the standard deviation of the variable, d is the margin of error (d was set as 5% of the mean value of the variable).

The body measurements which were used in the subsequent experiments were input as the variable to estimate the sample size for the three types of bodies. Table 1 shows the sample sizes calculated by primary body measurements.

Table 1.

The sample sizes of three types of bodies calculated by primary body measurements

Body measurement	CG _F	CG _B	WG _F	WG _B	SNW _F	SNW _B	SW
For Y-type bodies
SD (cm)	5.43	4.84	6.59	6.42	2.65	2.65	1.13
d (cm)	2.52	2.44	2.23	1.62	2.30	2.25	0.72
n	18	16	34	61	6	6	10
For A-type bodies
SD (cm)	4.69	4.77	4.90	6.20	1.97	1.99	1.04
d (cm)	2.41	2.31	2.21	1.68	2.23	2.19	0.69
n	15	17	19	53	3	4	9
For B-type bodies
SD (cm)	6.58	1.43	6.64	3.55	2.74	2.79	0.40
d (cm)	2.36	2.33	2.27	1.79	2.25	2.22	0.67
n	30	2	33	16	6	7	2

CG_F and CG_B are the front and back chest girth, respectively; and WG_F and WG_B are the front and back waist girth, respectively. SNW_F, SNW_B, and SW have been described previously. The detailed information of these body measurements will be explained in the subsequent context.

As shown in Table 1, the estimated least sample sizes are as follows: 61 for Y-type, 53 for A-type, and 33 for B-type. Therefore, 95 Y-type bodies and 55 A-type bodies which were selected before as the subjects are sufficient for the experiment. B-type bodies were excluded due to the small size. Further elementary measurements of the 150 subjects are shown in Table 2.

Table 2.

Elementary measurements of the subjects

Measurement item	Height (cm)	Weight (kg)	Body mass index (kg/m²)	Chest girth (cm)	Waist girth (cm)	Hip girth (cm)
Range	156.1–197.7	48–93.3	17.3–27.8	76.4–123.9	61.88–99.8	80.4–114.9
Mean	177.0	71.9	22.9	97.5	77.3	95.5
Standard deviation	8.2	8.1	2.0	7.9	7.2	5.7

As shown in Table 2, the measurements of the subjects distinctly varied in multiple dimensions. Most subjects were of normal stature (18.5 <body mass index (BMI) <25 kg/m²), and a few were skinny (BMI <18.5 kg/m²) or slightly obese (25 <BMI <30 kg/m²). Besides, several subjects who frequently had physical training were muscled with a BMI >25 kg/m².

To analyze body morphology, the scanner-compatible software Anthroscan (Human Solutions, Germany) was adopted to manipulate the scanning and to generate the 3D body mesh models. All these 3D models were visualized and measured by Rhinoceros (Robert McNeel & Associates, USA), the versatile 3D computer graphic and CAD software. For virtual simulation, the online platform Mixamo (Adobe, USA) was used to generate 3D boned scanatars (avatars transformed from a scanned body model) which were applicable in 3D CAD. For statistical analysis, SPSS (IBM, USA) software was used.

As a result, the regular untucked dress shirts (with classic forward point collar, turn round placket, round cuff, curved cut hem, and no back pleat) were customized based on the scanatars and their measurements. Moreover, four common shirt styles—body fit, slim fit, regular fit, loose fit—were involved with the variation in the volumes of bodice and the invariability in the neckline, collar, shoulder, cuff, sleeve length, and hem. 2D pattern blocks of these customized shirts were sketched in ET CAD (BUYI Technology, China), and the ultimate 3D try-on visualization was executed by CLO 3D (version 5.0.156.38765, CLO Virtual Fashion, Korea).

Acquisition of BM_N

As mentioned before, the BM_E for contemporary RtW apparel making such as H, CG, WG, HG, NG, SW, and AL were inadequate to characterize various individual morphological features. In our research, the BM_N were extracted from 3D mesh body models for characterizing the proportion and shape of the body and for controlling the balance of the individualized pattern block. To measure BM_N, the essential landmark points were labeled, such as SP, BNP (back neck point), SNP, FNP (front neck point), APF (armpit front fold point), APB (armpit back fold point), APD (deepest armpit point on axilla), EP (elbow point), and WP (wrist point) according to ISO 8559-1:2017(E).⁴⁰ Three critical levels—chest, waist, and hip—were marked as well.

Based on the landmarks, four groups of body measurements, namely heights, straight distances, circumferences, and surface lengths, were obtained in Rhinoceros by three methods: perpendicular-based, slice-based, and outline-based. Figure 1 shows the schemes of body measuring.

Figure 1.

Schemes of body measuring: (a) perpendicular-based, (b) slice-based, (c) outline-based, (d) divisional coronal plane.

As shown in Figure 1(a), the heights and straight distances were measured by the perpendicular (drawn by “Line” tool) from the landmarks to the ground and to the reference plane, respectively. Concretely, straight distances neck width (NW) and shoulder length (SL) were measured by directly connecting the bilateral landmarks.

As shown in Figure 1(b), the girths and surface lengths were measured from the cross-sections which were generated by intersecting the 3D body mesh with transverse and sagittal planes on different levels with “Mesh intersect” tool. Exceptionally, considering that FNP, SNPs, and BNP were not coplanar, front and back necklines were directly outlined on the mesh surface along the neck contour with the “Polyline on mesh” tool according to Figure 1(c). Hereby the front and back neck girths (NG_F and NG_B) were measured.

Moreover, the divisional coronal plane divided the torso girths (chest, waist, hip) into front and back parts according to Figure 1(d). This plane was made by connecting left SP and right SP and extending in the vertical direction with the “Rectangular plane: 3 points” tool. Hereby the girths were segmented into front and back parts by the divisional plane with the “Split” tool. According to the location of the body segment, the scheme and the list of BM_N and BM_E are shown in Figure 2 and Table 3.

Figure 2.

Scheme of measuring of BM_N and BM_E.

Table 3.

List of BM_N and BM_E

No.	Body segment	Body measurement		Abbreviation (Figure 2)	Definition
1	torso	BM _E	chest height	H _C	vertical distance from the chest level (axilla) to the ground
2			waist height	H _W	vertical distance from the waist level to the ground
3			hip height	H _H	vertical distance from the hip level to the ground
4			length from SNP to front waist	SNW _F	surface length from SNP to front waist line
5			length from SNP to back waist	SNW _B	surface length from SNP to back waist line
6			across front width	AW _F	surface length across the front between the left and the right arm scye lines at the level of APF
7			across back width	AW _B	surface length across the front between the left and the right arm scye lines at the level of APB
8		BM _N	front chest girth	CG _F	horizontal front part of chest girth measured at chest level
9			back chest girth	CG _B	horizontal back part of chest girth measured at chest level, CG_F + CG_B = CG
10			front waist girth	WG _F	horizontal front part of waist girth measured at waist level
11			back waist girth	WG _B	horizontal back part of waist girth measured at waist level, WG_F + WG_B = WG
12			front hip girth	HG _F	horizontal front part of hip girth measured at hip level
13			back hip girth	HG _B	horizontal back part of hip girth measured at hip level, HG_F + HG_B = HG
14	neck	BM _E	height of BNP	H _BNP	vertical distance from BNP to the ground
15		BM _N	height of FNP	H _FNP	vertical distance from FNP to the ground
16			height of SNP	H _SNP	vertical distance from SNP to the ground
17			neck width	NW	horizontal distance between left SNP and right SNP
18			horizontal depth of FNP	D _FNP	horizontal distance from FNP to predefined vertical plane
19			horizontal depth of SNP	D _SNP	horizontal distance from SNP to predefined vertical plane
20			horizontal depth of BNP	D _BNP	horizontal distance from BNP to predefined vertical plane
21			front neck girth	NG _F	front part of neck base girth
22			back neck girth	NG _B	back part of neck base girth, NG_F + NG_B = NG
23	shoulder	BM _E	shoulder length	SL	distance from SP to SNP
24		BM _N	length from SP to FNP	SSN _F	surface length from SP to FNP
25		BM _N	length from SP to BNP	SSN _B	surface length from SP to BNP
26	arm	BM _E	upper arm girth	UAG	upper arm girth measured through APD
27			arm length	AL	distance from SP to WP through EP
28			wrist girth	WRG	wrist girth at the level of WP

As shown in Figure 2 and Table 3, the morphological features of the torso, the neck, the shoulder, and the arm were supposed to be described by those body measurements. First, based on the torso measurements, not only the three horizontal levels and their girths but also the anteroposterior proportions of transverse girths and vertical surface lengths were expressed. Moreover, the positions of SNPs were fixed by SNW_F and SNW_B.

Second, the shape of neckline was identified by exact positions of FNP, BNP, and the anteroposterior proportion of neck girths. Third, for the shoulder, the 3D spatial positions of SPs were fixed accordingly by SL, SSN_F, and SSN_B, and the armscye lines were confirmed by AW_F and AW_B. Last, the three BM_E expressed the main features of the arm.

Generally, the whole upper body (including hip level) could be characterized by the combination of BM_N and BM_E with critical levels and landmarks exactly located and anteroposterior proportions correctly balanced.

Application of BM_N

Because BM_N and BM_E described the morphological features of the body, the method of using them to accomplish the customized shirt pattern directly has been proposed. For a customized e-bespoke shirt, the pattern construction should be associated with the corresponding body measurements.

The pattern block can be constructed stepwise by connecting referential points and using the pattern indexes which were calculated by equation (2)

I_{P - S} = BM \pm {E_{P -}}_{S}

(2)

where I_P–S is the value of pattern index, BM is the body measurement, E_P–S is the value of corresponding ease allowance, P represents the position of the pattern index, and S represents the style of shirt.

According to equation (2), the indexes I_P–S can assure not only the conformity to the individual body, but also the appropriate ease allowances for daily movements and activities. Exceptionally, some indexes equal BM with ease void since the ease addition will either have no influence on the apparel fit or result in misfit, for instance by SNW_F, SNW_B, SL, SSN_F, and SSN_B and the other converted BM_N proposed in the Results section.

For example, the pattern index I_CGF–S signifies the front width of the pattern on chest level which equals the sum of half front chest girth CG_F and ease to half front chest girth E_CGF

I_{CGF - S} = 0.5 C G_{F} + E_{CGF - S}

(3)

Table 4 explains the calculation of I_CGF–S for two scanned subjects S₃ and S₈₇, as examples (“S” stands for subject number).

Table 4.

Pattern index I_CGF–S

Subject No.	CG (cm)	CG_F (cm)	I_CGF–S (cm) for shirt style
Subject No.	CG (cm)	CG_F (cm)	Body fit	Slim fit	Regular fit	Loose fit
3	91.9	41.7	22.1	23.4	24.9	26.9
87	91.7	46.3	24.4	25.7	27.2	29.2

S₃ and S₈₇ belong to the same body type (170/92Y) with similar chest girth (91.9 and 91.7 cm respectively), but both CG_F values are different (41.7 and 46.3 cm respectively). Therefore, the customized patterns should be designed with different I_CGF–S values, as Table 4 shows. The values of E_CGF–S depend on shirt style, and the different CG_F values lead to different values of I_CGF–S.

Similarly, the other pattern indexes can be confirmed by corresponding body measurements and ease allowances.

Fit evaluation

We conducted the virtual try-on experiments in CLO 3D to evaluate the apparel fit offered by e-bespoke shirt patterns. Meanwhile, contemporary RtW and MtM patterns were also involved in verifying the difference between the three types of shirts.

For the consistency between real and virtual situations, three main elements were considered. First, we used the scanned bodies with individual morphology (scanatars). The scanatars with skeleton (format: .fbx) converted from raw scanning files (format: .obj) were imported into CLO 3D as the avatars. Second, the pattern blocks sketched by ET CAD were imported into CLO 3D and sewn as the real techniques. Third, the virtual fabrics were selected as the analogs of real materials with the same basic properties.

According to the aforementioned elements of accessing apparel fit and the common guidelines of men's shirt fitting, the subjective evaluation was conducted based on concrete criteria.

These criteria for a well-fitted shirt were the following: smooth surface without unnecessary folds and bulges; structure lines and points correspond to body anthropometrical levels and landmarks; balanced anteroposterior proportions. Chest, waist, and hip lines should be parallel to anthropometrical levels, side seams should be vertical, and there should be no folds and bulges brought forth by unbalanced proportions. The following five-point rating scale of fit was used: 1 = worst, 2 = poor, 3 = medium, 4 = good, 5 = best. In addition, the objective evaluation was concurrently conducted by comparing the strain maps built in CLO 3D. The strain map illustrates the garment distortion rate due to external stress by the range of color and numbers.

Textile materials

To achieve the virtual wearing effect that the real shirt has, a behavior of textile materials under shaping and draping should be taken into account. Five kinds of woven fabrics T₁–T₅ with different mechanical properties were selected for shirt sewing and try-on, and the properties were measured by KES-F (Japan).

As shown in Table 5, the chosen fabrics have disparate formability due to various properties. T₁ and T₂ are inelastic fabrics with light weight and thin thickness. T₃ and T₄ are thick hard fabrics with high stiffness. T₅ is the thickest fabric with high elasticity.

Table 5.

Properties of fabrics

Fabric	Basic information				Tensile						Shearing		Bending
	Fiber content	Weave structure	Thickness (cm)	Weight (g/m²)	EMT (%)		RT (%)		F (cN/cm)		G (cN/[cm · (°)])		B (cN · cm²/cm)
	Fiber content	Weave structure	Thickness (cm)	Weight (g/m²)	warp	weft	warp	weft	warp	weft	warp	weft	warp	weft
T ₁	100% polyester	satin	0.20	67.2	3.67	9.77	60.79	49.30	415.05	61.37	0.16	0.14	0.06	0.03
T ₂	100% cotton	plain	0.25	69.0	4.74	12.81	48.97	37.89	225.33	32.27	0.34	0.31	0.04	0.03
T ₃	100% cotton	plain	0.41	147.8	6.245	5.83	42.96	46.52	207.62	206.09	2.69	2.41	0.10	0.14
T ₄	98% cotton, 2% spandex	twill	0.41	183.0	5.37	7.07	41.93	52.82	273.88	130.68	3.00	2.09	0.00	0.25
T ₅	95% polyester, 5% spandex	satin	0.50	206.7	24.79	26.39	59.50	46.36	14.35	19.85	0.45	0.44	0.15	0.16

EMT is the elongation under 500 cN/cm, RT is the tensile resilience, F is the tensile strength when elongation is 3%, G is the shear rigidity, and B is the bending stiffness.

To import the fabrics' properties measured by KES-F to CLO 3D, a two-step adjustment with the virtual fabrics according to previous methods and algorithms was conducted.

First, the proximate virtual fabrics according to the basic information of real ones were selected with inputting the values of thickness and weight in CLO 3D. Second, we adapted the fabric properties (stretch-warp stiffness and stretch-weft stiffness, g/s²; shear, g/s²) in CLO 3D to real ones (EMT, RT, F, G) by the algorithms and sensory experiments.^41–43

The adjusted virtual fabrics show realistic wearing effect, helping us to precisely analyze and evaluate the apparel fit of simulated e-bespoke shirts.

Results

Pattern-oriented body measurements of neck and shoulder

Most of the body measurements can be directly applied to pattern customizing according to equation (2). However, some of the height and distance measurements were independent of pattern indexes. These measurements were converted into six BM_N for pattern drafting

D_{CW} = H_{C} - H_{W}

(4)

D_{WH} = H_{W} - H_{H}

(5)

D N_{F} = H_{SNP} - H_{FNP}

(6)

D N_{B} = H_{BNP} - H_{SNP}

(7)

NW L_{F} = D_{FNP} - D_{SNP}

(8)

NW L_{B} = D_{SNP} - D_{BNP}

(9)

where D_CW is the vertical depth from chest to waist, D_WH is the vertical depth from waist to hip, DN_F is the vertical depth from SNP to FNP, DN_B is the vertical depth from BNP to SNP, NWL_F is the lateral width of front neck, and NWL_B is the lateral width of back neck.

According to equations (4) to (9), D_CW and D_WH describe the vertical depths between chest, waist, and hip levels, which can help to draft the horizontal structure lines of the bodice.

DN_F, DN_B, NWL_F, and NWL_B depict spatial neck shape on the lateral view. In our approach, the neckline can be drafted from NG_F and NG_B, while the front drop and back drop are undefined with BM_E. The lateral projections of front and back necklines correspond to the neck drops; however, complex operations are necessary to measure the projections. To obtain the projections and neck drops, it was more convenient to calculate them with pre-existing BM_N immediately. Therefore, the neckline can be performed with new measurements (see Figure 3).

Figure 3.

Model of neckline and its transformation into pattern block: (a) neckline configuration and its projections of the scanatar; (b) neckline of the pattern block.

As shown in Figure 3(a), the neckline model in the 3D coordinate system can be presented by five landmarks: G (FNP), D (left SNP), E (right SNP), M (midpoint of line DE), and I (BNP). Front and back parts of the neckline were approximately located on rectangular planes MDHG and MDJI, respectively. The neckline and its landmarks were projected on three axial planes XOY, XOZ, and YOZ for showing front, lateral and top views. On YOZ plane, front and back necklines were projected as curve LPN_F (lateral projection of front neck) and curve LPN_B (lateral projection of back neck), respectively, which corresponded to the front and back drops (indexes I_LPNF and I_LPNB) as shown in Figure 3(b).

The vertical right triangles DHK and DJL, and their lateral projections D₃H₃K₃ and D₃J₃L₃ were drafted. From the multi-view of the neckline model, the equivalence relations were distinctly observed as equations (10) to (13). As a result, the lengths of hypotenuses HD and DJ were calculated in the right-angle triangles DHK and DJL as equations (14) and (15)

HK = H_{3} K_{3} = H_{1} D_{1} = NW L_{F}

(10)

DL = D_{3} L_{3} = D_{1} J_{1} = NW L_{B}

(11)

DK = D_{3} K_{3} = D_{2} H_{2} = D N_{F}

(12)

JL = J_{3} L_{3} = D_{2} J_{2} = D N_{B}

(13)

HD = \sqrt{D K^{2} + H K^{2}} = \sqrt{D {N_{F}}^{2} + NW {L_{F}}^{2}}

(14)

DJ = \sqrt{J L^{2} + D L^{2}} = \sqrt{D {N_{B}}^{2} + NW {L_{B}}^{2}}

(15)

where HK, H₃K₃, and H₁D₁ are the lateral widths of front neck; DL, D₃L₃, and D₁J₁ are the lateral widths of back neck; DK, D₃K₃, and D₂H₂ are the vertical depths from SNP to FNP; JL, J₃L₃, and D₂J₂ are the vertical depths from BNP to SNP; HD and DJ are the straight line segments that approximated LPN_F and LPN_B, respectively.

As shown in Table 6, after comparison of projections LPN_F and LPN_B and approximate hypotenuses HD and DJ of 150 subjects, the maximum absolute error was up to 0.61 cm, which would affect the neck fitting for some bodies. To improve the prediction model, we proposed the next indexes LPN_F–P and LPN_B–P (predicted LPN_F and LPN_B) to estimate the projection values, linearly expressed with HD and DJ.

Table 6.

Absolute errors of predicted front and back lateral neck projections

Absolute error	Mean (cm)	Range (cm)
\| LPN_F – HD\|	0.24	0.05–0.61
\| LPN_B – DJ\|	0.06	0.06–0.25
\| LPN_F – LPN_F–P \|	0.08	0.00–0.21
\| LPN_B – LPN_B–P \|	0.03	0.00–0.19

Values of LPN_F, LPN_B, HD, and DJ of 150 subjects were imported in SPSS, and their linear relations were analyzed. LPN_F–P and LPN_B–P can be calculated by linear regression equations (16) and (17) respectively

\begin{matrix} LP N_{F - P} = 1.057 HD - 0.322 \\ 1.057 \sqrt{D {N_{F}}^{2} + NW {L_{F}}^{2}} - 0.322 \end{matrix}

(16)

\begin{matrix} LP N_{B - P} = 0.994 DJ + 0.076 \end{matrix} 0.994 \sqrt{D {N_{B}}^{2} + NW {L_{B}}^{2}} + 0.076

(17)

The correlation coefficients calculated were 0.998 between LPN_F–P and HD, 0.999 between LPN_B–P and DJ, and both coefficients were significant at 0.01 level (two-tailed). Table 6 shows the accuracy of both indexes LPN_F–P and LPN_B–P. The mean and maximum absolute errors decreased to a lower level, which is acceptable for individual neck drop drafting.

Therefore, utilizing five pattern-oriented BM_N variables—NW, LPN_F–P, LPN_B–P, NG_F, and NG_B— the neckline can be exactly drafted with corresponding indexes I_NW, I_LPNF, I_LPNB, I_NGF, and I_NGB respectively, with FNP, SNP, and BNP correctly positioned.

During pattern drafting, shoulder lines are configurated by the bilateral points SNP and SP. SNP can be fixed by neck measurements. To locate SP, three measurements—SL, SSN_F, and SSN_B—were involved in transforming 3D lines on the scanatar into 2D lines of the pattern (see Figure 4).

Figure 4.

Model of shoulder area and scheme of its flattening: (a) shoulder area of the scanatar; (b) shoulder lines of the pattern block.

As shown in Figure 4(a), SL determined shoulder length in the pattern, and SSN_F and SSN_B reflected shoulder sloping and shoulder direction (forward thrust, back thrust, etc.). The 2D pattern can be obtained by intersecting three circles with corresponding radius indexes I_SL, I_SSNF, and I_SSNB, and center points SNP, FNP, and BNP, as shown in Figure 4(b). In this manner, the shoulder line of the pattern will coincide with the body shoulder area.

Pattern-oriented body measurements of torso and arm

The bodice of a men's shirt is the most decisive part of the overall apparel fit. In this section, primary problems of misfits relating to the unbalanced anteroposterior proportions, which influence the smoothness of the surface and the straightness of structure lines, were solved. Figure 5(a) and (b) show the model of the male torso and the scheme of the bodice pattern.

Figure 5.

Models of torso and arm, and their transformation into flat patterns of bodice and sleeve: (a) torso and arm of scanatar; (b) bodice pattern block; (c) sleeve pattern block.

As shown in Figure 5(a) and (b), D_CW and D_WH and corresponding pattern indexes I_DCW and I_DWH helped to find chest, waist, and hip levels. The six BM_N variables CG_F, CG_B, WG_F, WG_B, HG_F, and HG_B and their corresponding pattern indexes I_CGF, I_CGB, I_WGF, I_WGB, I_HGF, and I_HGB determined the horizontal anteroposterior bodice proportion by front and back parts of girths. Stress folds and tilt side seams will take place when the proportion is unbalanced. SNW_F and SNW_B and pattern indexes I_SNWF and I_SNWB controlled the vertical anteroposterior proportion by connecting SNP and waist level on the front and the back. The chest, waist, and hip structure lines will be tilted when this proportion is unbalanced. After finding APD and SP by the aforementioned measurements, AW_F and AW_B and pattern indexes I_AWF and I_AWB helped to draft the armhole curves.

Figure 5(a) and (c) shows the measurements AL, UAG, and WRG and sleeve indexes I_AL, I_UAG, and I_WRG which are commonly used to exactly configurate the sleeve length, sleeve width, and cuff width, respectively.

With 23 converted and unconverted body measurements, the points and the lines of pattern block can be arranged accordingly. Each part of the shirt pattern block can be accomplished by body measurements of different body sections to produce the e-bespoke shirt.

Customized shirt pattern block

BM_E and BM_N could be applied directly to construct the customized pattern block by stepwise connecting referential points with a predefined ease allowance addition. Figure 6 shows the scheme of customizing the shirt pattern.

Figure 6.

Scheme of customizing the pattern block of a men's shirt.

As shown in Figure 6, the shirt pattern block was customized by connecting 39 points in sequence with pattern indexes which were calculated from corresponding body measurements and ease allowances using equation (2).

There were three kinds of ease allowances applied: zero, constant, and variable. First, some of the indexes should equal the raw body measurements without ease addition for good fitting, and these are I_SL, I_SSNF, I_SSNB, I_SNWF, I_SNWB, I_DCW, and I_DWH. Second, some of the indexes involved eases which were invariable with style change. Lastly, some of the indexes should be variable as the shirt style changed. Table 7 shows the second and third kinds of ease allowances.

Table 7.

Ease allowances for different shirt styles

Ease allowance (cm)	Shirt style
Ease allowance (cm)	Body fit	Slim fit	Regular fit	Loose fit
to neck width E_NW	1.0
to lateral projection of front neck E_LPNF	0.5
to lateral projection of back neck E_LPNB	−0.5
to front neck girth E_NGF	1.5
to back neck girth E_NGB	0.5
to arm length E_AL	3.0
to wrist girth E_WRG	6.0
to chest girth E_CG	5.0	10.0	16.0	24.0
to waist girth E_WG	14.0	18.0	28.0	40.0
to hip girth E_HG	7.0	9.0	12.0	20.0
to axilla depth E_AD	3.0	4.5	6.0	7.5
to upper arm girth E_UAG	10.0	12.0	15.0	19.0
to across front width E_AWF	2.0	4.0	6.0	8.0
to across back width E_AWB	2.0	4.0	6.0	8.0

As shown in Table 7, E_LPNB was an exceptionally negative value, signifying shortening back neck drop. When there were positive E_NW and E_NGB, the neckline would fall down below BNP because of the oversized back neckline. To offset the falling-down effect, the back neck drop was shortened to uplift the back neckline to the correct position.

To illustrate the influence of BM_N and the ability of the new database created to present the morphological features of male bodies, the customized patterns of shirts in body fit style were designed for seven subjects of 170/92Y type and overlapped. These subjects were identified as S₃, S₈, S₃₄, S₃₇, S₅₆, S₇₂, and S₈₇. Although the subjects belong to the same body type, their BM_N are distinctly different. For instance, the proportions between the parts of chest girth are disparate among subjects S₅₆ (CG_F = 39.1 cm, CG_B = 52.7 cm), S₇₂ (CG_F = 53.2 cm, CG_B = 40.7 cm), and S₈₇ (CG_F = 46.3 cm, CG_B = 45.4 cm). Figure 7 shows the overlapped RtW, MtM, and e-bespoke shirt patterns of body fit style. As usual, the armhole dart is only applied in women's wear. However, the front breast of S₃₇ and S₇₂ was too bulgy, and the armhole darts were applied to make the scye and chest area smooth.

Figure 7.

Pattern blocks: (a) RtW and e-bespoke patterns for 170/92Y scanatars; (b) RtW, MtM, and e-bespoke bodice patterns for subject S₇₂.

The RtW pattern was drafted based on the seven BM_E which were taken from the sizing system (H, CG, WG, HG, NG, SW, AL). As shown in Figure 7(a), significant differences not only existed between the two categories of RtW and e-bespoke, but also existed between all e-bespoke patterns in the proportions of the bodice, neckline, shoulder line, sleeve, and cuff.

The MtM pattern was drafted by adapting the RtW pattern to individual bodies by means of 16 BM_E (CG, WG, HG, NG, SL, AL, WRG, UAG, SNW_F, SNW_B, H_B, H_W, H_H, H_BNP, AW_F, AW_B) from the contemporary researches and MtM services. Figure 7(b) shows three types of bodice patterns for subject S₇₂ as an example: sleeves and collar were not shown due to the same MtM and e-bespoke patterns. Compared with the RtW pattern, the MtM pattern modified the proportion of vertical lengths (from SNP to anteroposterior waist), neck girth, and shoulder length with the next BM_E. e-Bespoke and MtM patterns were similar in the proportion of vertical lengths, shoulder length, and neck girth. However, the anteroposterior proportions of bodice, neckline, and the location of shoulder line showed a big difference owing to BM_N.

To illustrate the influence of BM_N on the different shirt styles, four pattern blocks which were designed for the subject S₇₂ were overlapped (see Figure 8). The patterns of collar and cuff are not shown because they are invariable with the style.

Figure 8.

Pattern blocks of four shirt styles for subject S₇₂.

As shown in Figure 8, the configuration of customized pattern blocks (bodice, sleeve) was escalating from body fit style to loose fit style with incremental ease allowances on bodice and sleeve.

Fit evaluation of e-bespoke shirts

To check the proposed body measurements and e-bespoke pattern blocks, subjective and objective evaluations were conducted. For subjective evaluation, the experts in apparel design were enrolled to subjectively evaluate the virtual images generated by CLO 3D. They were required to learn and to rate the wearing effects of e-bespoke shirts using a five-point scale in consideration of smoothness, locations of structure lines and landmarks, and anteroposterior balances. For objective evaluation, the built-in strain map module was utilized to visualize and analyze the variance of stress tension of the shirts in CLO 3D.

Based on the combination of 23 converted and unconverted BM_N and BM_E, and corresponding pattern drafting method, the e-bespoke shirts were simulated in CLO 3D. Among the shirts in four different styles, body fit shirts tend to exhibit the most misfits due to the thinnest air gaps between the fabrics and the bodies, which are configurated by the body measurements and ease allowances. When the configuration is incongruent with the morphology and proportion of the scanatar, the evident misfit can be observed. Figure 9 shows the exterior appearances of e-bespoke body fit shirts on seven 170/92Y subjects (S₃, S₈, S₃₄, S₃₇, S₅₆, S₇₂, S₈₇).

Figure 9.

Try-on effects of e-bespoke shirt in body fit style for the 170/92Y scanatars front and profile holistic appearances: (a) S₃; (b) S₈; (c) S₃₄; (d) S₃₇; (e) S₅₆; (f) S₇₂; (g) S₈₇.

As shown in Figure 9, the black lines are the primary landmarks (bust, waist, hip, FNP, SNP, shoulder line) on the surface of scanatars, and the red lines are the structural lines of the shirts. Seven e-bespoke shirts exhibit good apparel fit despite the significantly different customized patterns. Overall, the torso, shoulder, and armhole segments are smooth; chestlines, waistlines, and hiplines are horizontally located near the natural levels; side seams are vertical to the ground; necklines and shoulder lines precisely coincide with the landmarks. Concretely, the diagonal fold exists on the profile of S₅₆ as a result of an overlarge back proportion (CG_F = 39.1 cm, CG_B = 52.7 cm, HG_F = 38.7 cm, HG_B = 50.4 cm), and the slight bulge on the front of S₈₇ is due to an overlarge front proportion (WG_F = 41.3 cm, WG_B = 30.7 cm, HG_F = 51.9 cm, HG_B = 42.9 cm). Those whose shoulder is forward thrust (S₃, S₅₆, S₈₇) tend to show a slightly unsmooth armhole with pointed bulges. Moreover, the yoke cloth cannot cling to the shoulder due to the concave mid shoulders of subjects S₈, S₃₄, S₃₇, and S₅₆. Table 8 shows the scores of the subjective evaluation of seven e-bespoke shirts for 170/92Y subjects.

Table 8.

Subjective evaluation of e-bespoke shirts in body fit style

Subject	Evaluation of shirt section				Overall score
Subject	Bodice	Neck, collar	Shoulder	Armhole, sleeve	Overall score
S ₃	5	4	3	4	4
S ₈	4	5	5	5	5
S ₃₄	5	4	4	4	4
S ₃₇	4	5	4	5	5
S ₅₆	4	5	4	4	4
S ₇₂	4	5	5	4	5
S ₈₇	4	5	4	4	4

As shown in Table 8, all e-bespoke shirts were rated as good apparel fit with overall scores ranging from 4 to 5. For most subjects, scores of bodice, neck, shoulder, and sleeve were ≥4 points. Exceptionally, the shoulder of S₃ was evaluated as the medium fit due to the minor asymmetry of left and right shoulder slopings.

Comparison between RtW, MtM, and e-bespoke shirts

Comparative virtual try-on experiments of shirts produced by RtW, contemporary MtM customization, and e-bespoke in CLO 3D, respectively, were conducted. To concretely illustrate the difference between three kinds of shirts, the case of subject S₇₂ was exhibited as an example. Figure 10 shows the virtual try-on effects of RtW and MtM shirts on S₇₂.

Figure 10.

Front, back, and profile views of torso, and front view of neck and shoulder of try-on effects in body fit style for subject S₇₂: (a) RtW; (b) MtM; (c) e-bespoke.

As shown in Figure 10, the RtW shirt exhibits an evident poor apparel fit with distortion in the back and armhole, tight neckline, and too long shoulder and sleeve. With more individual BM_E applied, the MtM shirt is better fitted to the scanatar with smoother back contour and armhole, looser neckline, and shorter shoulder and sleeve. However, some misfit features still remain in the MtM shirt: oblique side seam, bust, waist and hip lines, malposition of landmarks (FNP, BNP, SNP, SP), and big bulge on SP. The reason is that BM_E still cannot characterize the primary body morphology. First, NG alone cannot well reflect the concrete neckline shape and landmarks. Second, SW alone cannot precisely locate the shoulder and SP. Third, the anteroposterior proportion of the bodice cannot be balanced with only CG, WG, and HG. Although SNW_F and SNW_B helped to balance the proportion of vertical lengths, it would also be implicated by the unbalanced girth proportions.

The e-bespoke resolved the misfit appeared in the RtW and MtM shirts by involving the next BM_N which were measured at the new position of the body. Moreover, the new pattern indexes were confirmed by adding BM_N and new ease values together as in equation (2). In reverse, the new ease allowance value can be calculated by pattern indexes minus BM_N as in equation (18)

E_{PN} = I_{PN} - B M_{N}

(18)

where E_PN is the value of new ease allowance, I_PN is the value of new pattern index, and PN represents the position of the pattern index corresponding to the BM_N.

For e-bespoke shirts, E_PN was predefined with concrete values to ensure adequate volume for particular body segments. However, BM_N were not adopted by the RtW and MtM shirts, and consequently E_PN and corresponding I_PN were unknown during pattern drafting. For example, NG and E_NG (ease to neck girth) are adopted to draft the necklines of RtW and MtM patterns, but NG_F and E_NGF (ease to front neck girth) are still unknown. To analyze the reason for misfit in RtW and MtM shirts and explain the impact of BM_N on e-bespoke patterns, E_PN of three types of shirts was compared using equation (19)

E_{PN} = P M_{PN} - B M_{N}

(19)

where PM_PN is the value of the new pattern measurement that corresponds to the position of BM_N.

PM_PN exactly equaled I_PN in the e-bespoke patterns, while PM_PN were posteriorly measured after RtW and MtM patterns completed by BM_E. Table 7 shows BM_N of S₇₂ and E_PN and PM_PN of three types of patterns. E_PN of e-bespoke patterns were predefined. However, E_PN of RtW and MtM patterns were varied: too small values, too large values, or acceptable values.

As shown in Table 9 and Figure 10, the ease values E_PN explained the misfit of the RtW and MtM shirts. For the RtW pattern, except for E_NW, the other four eases (E_LPNF, E_LPNB, E_NGF, E_NGB) were negative (pattern measurements were smaller than body measurements), resulting in the small neckline above natural FNP and BNP. For the MtM pattern, large E_NW and E_NGB led to the profile and back necklines lower than natural SNP and FNP, and negative E_LPNF and E_NGF led to the upper front neckline. Both of the two types of patterns neglected the proportion of front and back neck. Moreover, excessive E_SSNF and E_SSNB gave rise to the improper shoulder sloping and SP location of the shirts with the bulge above the shoulder. Lastly, inadequate front girths of the pattern with negative E_CGF, E_WGF, and E_HGF resulted in the stressed folds and oblique structural lines on the bodices. We simultaneously utilized the strain map of three types of shirts in CLO 3D to show the impact of BM_N. The turndown collars were concealed for better observation of necklines, and sleeves were concealed because BM_E made the entire pattern.

Table 9.

BM_N of subject S₇₂, and PM_PN and E_PN of RtW, MtM, and e-bespoke patterns

BM _N	Item		NW	LPN _F	LPN _B	NG _F	NG _B	SSN _F	SSN _B	CG _F	CG _B	WG _F	WG _B	HG _F	HG _B
BM _N	Value (cm)		13.6	9.6	3.2	25.6	15.0	21.4	20.2	53.2	40.7	45.2	28.2	49.9	41.5
PM _PN	Item		PM _NW	PM _LPNF	PM _LPNB	PM _NGF	PM _NGB	PM _SSNF	PM _SSNB	PM _CGF	PM _CGB	PM _WGF	PM _WGB	PM _HGF	PM _HGB
	Value (cm)	RtW	14.4	8.8	1.4	25.2	13.9	23.0	22.2	48.5	48.5	44.0	44.0	50.1	50.1
		MtM	15.7	8.8	2.7	25.4	17.2	21.7	21.1	49.5	49.5	43.8	43.8	49.2	49.2
		e-bespoke	14.6	10.1	2.7	27.1	15.5	21.4	20.2	55.7	43.2	50.0	37.5	55.5	43.0
E _PN	Item		E _NW	E _LPNF	E _LPNB	E _NGF	E _NGB	E _SSNF	E _SSNB	E _CGF	E _CGB	E _WGF	E _WGB	E _HGF	E _HGB
	Value (cm)	RtW	0.8	−0.8	−1.8	−0.4	−1.1	1.6	2.0	−4.7	7.8	−1.2	13.8	0.2	8.6
		MtM	2.1	−0.8	−0.5	−0.2	2.2	0.3	0.9	−3.7	8.8	−1.4	13.6	−0.7	7.7
		e-bespoke	1.0	0.5	−0.5	1.5	0.5	0	0	2.5	2.5	4.8	9.2	5.5	1.5

Figure 11(a) shows the strain maps of three types of shirts, which visualize the garment's distortion rate due to external stress indicated by different colors.

Figure 11.

Strain maps of RtW, MtM, and e-bespoke shirts in body fit style on subject S₇₂: (a) strain maps; (b) segments for measuring the distortion rate.

Figure 11(b) and Figure 12(a) show the distortion rates of different segments of three types of shirts. As shown in Figure 11(b), the shirt bodice was segmented in to 15 parts to demonstrate how BM_N affected the distortion of different areas.

Figure 12.

Variation of distortion rates of RtW, MtM, and e-bespoke shirts in body fit style: (a) subject S₇₂; (b) mean values of seven 170/92Y subjects.

As we can see, e-bespoke exhibited the least distortion area. The asymmetric neck, the large front chest, and the prominent tissue on the front waist led to the relatively higher distortion rates in the left neck, second button on the placket, and front waistline, respectively. By contrast, the MtM shirt showed more distortion in front chest and waist, scapula, and back armhole owing to the stress force generated by the negative eases of the front torso. Moreover, the front and profile neck segments were distorted the least, not because of the fitted neckline, but because of the loose neck width and unfitted neckline distant from natural FNP and SNP. The whole body of the RtW shirt was distorted from neck to the hem because of the too small neckline and front girths. The reason for less distortion in the SP area was the excessive up shoulder sloping and long shoulder width (shoulder line and natural shoulder were not contiguous with the big bulge around according to Figure 10(a)). Similarly, the unfitted excessive back armhole led to less distortion in the back armhole area.

In the same way, the distortion rates of virtual RtW, MtM, and e-bespoke body fit shirts for seven 170/92Y scanatars (in total 21 pieces) were measured in CLO 3D. As shown in Figure 12(b), most segments of e-bespoke shirts were distorted the least among the three types. As usual, the collar should surround the neck through landmarks, and the whole shirt should overhang on the torso above the chest level (from the upper chest, shoulder, to scapula) with inevitable distortion on the cloth. Exceptionally, the lowest rate in N₁ of MtM shirts was caused by the unfitted necklines upper than the natural necks of most subjects (as Figure 10(b)), the lower rate in A₁ of RtW and MtM shirts was caused by the excessive up shoulder sloping of most subjects (as Figure 10(a) and (b)), and the lowest rate in B₁ of RtW shirts was caused by the malposition of front and back piece (more strain on the front and less strain on the back). In accordance with the aforementioned criteria of fit evaluation and the five-point rating scale, the subjective evaluation of the RtW and MtM shirts for the seven subjects was simultaneously executed by the experts. Table 10 shows the mean scores of the subjective evaluation of three types of shirts.

Table 10.

Mean scores of subjective evaluations of RtW, MtM, and e-bespoke shirts in body fit style

Shirt type	Evaluation of shirt section				Overall score
Shirt type	Bodice	Neck, collar	Shoulder	Armhole, sleeve	Overall score
RtW	1.3	1.3	1.1	1.6	1.4
MtM	2.6	1.6	2.1	2.9	2.6
e-bespoke	4.1	4.7	4.1	4.3	4.4

As shown in Table 10, RtW, MtM, and e-bespoke shirts of body fit style were generally rated as worst fit, medium fit, and good fit levels, respectively. Although MtM shirts were made by the individual BM_E, the apparel fit obtained was not satisfactory, especially in bodice, neck, and shoulder. By contrast, the e-bespoke exhibited good apparel fit of every segment with the next BM_N and new customized pattern involved. Therefore, from both results of subjective and objective evaluation of three types of shirts, the e-bespoke men's shirt can assure the invariable good fit for the Y-type and A-type subjects, which is superior to the contemporary RtW and MtM shirt.

e-Bespoke shirts in different styles

Shirt patterns of the other three styles (slim fit, regular fit, loose fit) were verified simultaneously in CLO 3D. As an example, the results of S₇₂ are shown in Figure 13.

Figure 13.

Try-on effects of e-bespoke shirts for subject S₇₂: (a) slim fit; (b) regular fit; (c) loose fit.

As shown in Figure 13, for every kind of style, the shirts exhibited overall good apparel fit. From body fit style to loose fit style, the amount of stressed folds (misfit) decreased, the amount of natural draping folds increased, and the volume of bodice and sleeve became bigger. Meanwhile, the positions of necklines, shoulder lines, and cuff hems were invariable due to the constant ease allowances in the corresponding pattern sections. Exceptionally, the bodice pieces were not vertical due to the slightly slant torso. The huge volumes of bodices and sleeves of regular fit and loose fit shirts were inevitably collided, giving rise to the slightly deformed armholes.

Therefore, the combination of BM_E and BM_N, as well as the corresponding customized pattern, can be applied to different shirt styles from body fit to loose fit with assured high-level apparel fit.

e-Bespoke shirts made of different fabrics

We applied five aforementioned fabrics to the same body fit shirt for subject S₇₂. Due to the close properties of fabric T₁ and T₂, the shirts made of T₁ and T₂ exhibited similar appearances and fit. The shirts made of T₃ and T₄ were of similar appearance likewise. Figure 14 shows the representative appearances of virtual shirts made of fabrics T₁, T₄, and T₅.

Figure 14.

Try-on effects of shirts made from different fabrics: (a) T₁; (b) T₄; (c) T₅.

As shown in Figure 14, some tiny differences are observable in the three shirts. Fabric T₁ is inelastic and soft, and some tiny folds exist on the front chest and the back of the shirt. T₅ is of high elasticity and low rigidity, and the shirt shows less folds and a smoother surface. T₄ is of high shearing rigidity, and the shirt shows a starched appearance with fewest folds. Nonetheless, the structure lines of all the three shirts are well-located and the proportions are balanced.

Therefore, the proposed body measurements and customized pattern can also assure a similar good fit for shirts made from different fabrics.

Discussion

For the improvement of apparel fit of customized men's shirts, the BM_N and pattern drafting method should be utilized to characterize and match the individual morphological features. The results obtained indicate that the virtual e-bespoke shirts based on BM_N and new customized patterns can assure the high-level apparel fit, which is available for different bodies, shirt style, and textile materials.

The contemporary RtW shirt pattern commonly utilized seven BM_E from the sizing system to fit the typical bodies, and was the worst apparel fit consequently shown on the various bodies. For improvement of apparel fit, the contemporary MtM shirt adapts the RtW pattern to the 16 BM_E to characterize in detail the individual morphology of torso and arm. However, the malposition of the neckline, shoulder line, and the unbalanced proportion of torso can also result in poor and medium apparel fit. Moreover, conventional RtW and MtM shirts required a repetitive real try-on process for adjustment. With the involvement of 3D scanning technology and 3D CAD, the e-bespoke patterns can be directly drafted through the new combination of 28 raw BM_E and BM_N, which can assure a good and best apparel fit without real try-on. Compared to BM_E utilized in RtW and MtM, BM_N promotes the customized pattern with three new variants. First, the torso variant further emphasizes the anteroposterior proportions in both aspects of girths and body lengths to ensure the proper allocation of front and back volumes and exact position of structural lines. Second, the neck variant comprehensively depicts the 3D shape and proportion of the neckline to make the correct location of neckline, collar, and the corresponding feature points (FNP, SNP, BNP). Third, the shoulder variant can accurately locate SP not only with real shoulder width but also with direction and sloping of real shoulder. Moreover, the corresponding pattern drafting method helps to systematically assemble each segment into one integrated pattern that fits the individual morphology.

Although the proposed e-bespoke shirts assure good apparel fit on different subjects, the generalizability of the results is limited by some factors. For example, the Y and A body types were selected as more typical for young men as the subjects in consideration of the validity of sample size; the proposed approach is unproved for the obese population (B and C types). Moreover, the usual scenario of men's shirt dressing (untucked shirt, static standing pose) has been investigated, but more experiments and analyses are necessary for the other scenarios, for instance tucked shirt and active poses. In the future, developed approaches will be optimized with the obese population enrolled.

The additional experiments of comprehensive scenarios will be performed with both real and virtual samples to validate the practicability of developed approach.

In summary, the new virtual e-bespoke men's shirt can assure the good individual apparel fit based on the BM_N and customized pattern, which is superior to the contemporary RtW and MtM methods. The proposed e-bespoke approach has the industrial implication for shirt customization services by the transformation from RtW patterns or directly drafting the latest pattern with the BM_N. The geometric models of torso, neck, and shoulder can be extended to the customization of other categories of men's multilayer apparel such as jacket, coat, and so on.

Conclusion

This paper presented the new method of customizing a virtual e-bespoke men's shirt based on BM_N and corresponding method of pattern drafting. With the 150 Y-type and A-type scanatars, 28 raw body measurements were measured in 3D modeling software. At the same time, the geometric models of male torso, neck, shoulder, arm, and the method of applying the models to pattern customization were established for characterizing different body segments with the combination of 23 converted and unconverted BM_E and BM_N. The virtual e-bespoke shirts show good apparel fit in both subjective and objective evaluations with balanced proportion, correct location of structural lines and feature points, and low fabric strain. By contrast, the contemporary RtW and MtM shirts only obtained poor fit and medium fit, respectively. The e-bespoke shirt was also proven feasible for different styles and textile materials.

The proposed BM_N and pattern drafting method are recommended to extend to the process of customization to produce well-fitted men's shirts without repetitive real try-on tests. The geometric models can also be extended to the customization of other categories of men's apparel. In the future, more body types and scenarios will be performed with further real and virtual experiments. The digital twin for e-bespoke men's shirts will be established by adopting the database of body morphology, pattern customization, textile material, and so on.

Footnotes

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 support for the research, authorship, and/or publication of this article: This study was supported by the Russian Ministry of Science and Education under project number RFMEFI61619X0113 (05.616.21.0113).

ORCID iD

Victor E Kuzmichev

References

Amed I, Andersson J, Berg A, et al.The state of fashion 2018. McKinsey & Company and The Business of Fashion, 2017, pp. 44–45.

China Electronic Commerce Research Center. 2015–2016 China garment e-commerce industry report. China Electronic Commerce Research Center, 2016, pp. 12–25.

Ministry of Industry and Information Technology. Development plan of textile industry (2016–2020). Ministry of Industry and Information Technology, 2016.

China National Garment Association. 2016–2017 China garment industry development report. 1st ed. Beijing: China Textile & Apparel Press, 2017, pp. 84–90.

KPMG and AliResearch. From tool-driven revolution to decision-making revolution: a path to smart manufacturing transformation. KPMG and AliResearch Institute, 2017, p. 5.

Compilation and Publishing Center of Changshu Men's Wear Index of China. Report on the development of China men's wear industry in 2017. 1st ed. Beijing: China Textile & Apparel Press, 2017, pp. 173–177.

Alexander

Connell

Presley

. Clothing fit preferences of young female adult consumers. Int J Cloth Sci Tech 2005; 17: 52–64.

Surikova

Kuzmichev

Surikova

. Improvement of clothes fit for different female bodies. Autex Res J 2017; 17: 111–119.

Gill

. Improving garment fit and function through ease quantification. J Fash Mark Manag 2011; 15: 228–241.

10.

Erwin

Kinchen

Peters

. Clothing for moderns, 6th ed. New York: Macmillan Publishing, 1979.

11.

GB/T 22187-2008. General requirements for establishing anthropometric databases.

12.

ISO 15535:2006. General requirements for establishing anthropometric databases.

13.

Winks

. Clothing sizes: international standardization, Manchester: Textile Institute, 1997.

14.

Anderson LJ, Brannon EL, Ulrich PV, et al. Understanding fitting preferences of female consumers: development of an expert system to enhance accurate sizing selection. National Textile Center Annual Report, 2001, 198, pp.1–10.

15.

Brackelsberg P and Marshall R. Unit method of clothing construction. Waveland Press Inc, 1990.

16.

Song G. Improving comfort in clothing. : Woodhead Publishing, 2011, pp. 245-277.

17.

Brooks Brothers. Men dress shirts size & fit guides, https://www.brooksbrothers.com/on/demandware.store/Sites-brooksbrothers-Site/ru_RU/Page-FitSizeGuide?guideID=men-dress-shirts&activeTab=fit (2019, accessed 4 August 2019).

18.

Budd Shirtmakers. Made to measure shirts, https://www.buddshirts.co.uk/shirts/made-to-measure (2019, accessed 30 July 2019).

19.

Bivolino.com. Made to measure shirts—designer—3D shirt configurator—create your own!, https://www.bivolino.com/en/made-to-measure-shirts (2019, accessed 30 July 2019).

20.

Uniqlo Co. Ltd. Customized for you: shop easy care shirts, 2019. Available at: https://www.uniqlo.com/us/en/page/customizedforyou-easycareshirts-fit.html (accessed 30 July 2019).

21.

Spier & Mackay. Custom made to measure shirts online, https://www.spierandmackay.com/design-your-custom-shirt.html (2018, accessed 30 July 2019).

22.

Proper Cloth. Shirt sizes, https://propercloth.com/sizes/measure-body (2019, accessed 30 July 2019).

23.

iTailor. Custom dress shirt, 2019. Available at: https://www.itailor.com/designshirts/ (accessed 30 July 2019).

24.

Tailor Store Sweden. Tailor store, https://www.tailorstore.com/mens-body-measurements (2019, accessed 30 July 2019).

25.

MatchU Tailor. Indigo triangle houndstooth cotton, https://www.matchutailor.com/collections/matchu-dress-shirts/products/indigo-triangle-houndstooth-cotton (2019, accessed 30 July 2019).

26.

Indochino. Custom suits & men's suits, https://www.indochino.com/measurements (2019, accessed 30 July 2019).

27.

Istook

Hwang

. 3D body scanning systems with application to the apparel industry. J Fash Mark Manag 2001; 5: 120–132.

28.

Gupta D and Zakaria N. Anthropometry, apparel sizing and design. Cambridge: Woodhead Publishing, 2014, pp. 305–319.

29.

Hong

Zeng

Bruniaux

, et al. Interactive virtual try-on based three-dimensional garment block design for disabled people of scoliosis type. Text Res J 2016; 87: 1261–1274.

30.

Lee

Hong

Lee

. Development of 3D patterns for functional outdoor pants based on skin length deformation during movement. Int J Cloth Sci Tech 2017; 29: 148–165.

31.

Chan

Fan

. Men's shirt pattern design. Part I: an experimental evaluation of shirt pattern drafting methods. Sen'i Gakkaishi 2003; 59: 319–327.

32.

Liu

Zhu

Tao

, et al. Parametric design of garment pattern based on body dimensions. Int J Ind Ergonom 2019; 72: 212–221.

33.

Liu

. Individualizing women's suit patterns using body measurements from two-dimensional images. Text Res J 2017; 87: 669–681.

34.

Liu

Wang

Kamalha

, et al. Construction of a prediction model for body dimensions used in garment pattern making based on anthropometric data learning. J Text Inst 2017; 108: 2107–2114.

35.

Kim

Innami

Takatera

, et al. Individualized male dress shirt adjustments using a novel method for measuring shoulder shape. Int J Cloth Sci Tech 2017; 29: 215–225.

36.

Han

. Investigation of chest girth and height variation affecting other body dimensions for men's upper body clothes pattern grading. Int J Cloth Sci Tech 2019; 31: 415–430.

37.

Jang

Chang

. A study on a men's dress shirt pattern by somatotype for mass customization system. J Korean Soc Cloth Text 2008; 32: 294–306.

38.

ISO 20685:2010(E). 3-D scanning methodologies for internationally compatible anthropometric databases.

39.

Charan

Biswas

. How to calculate sample size for different study designs in medical research?. Indian J Psychol Med 2013; 35: 121–126.

40.

ISO 8559-1:2017(E). Size designation of clothes—part 1: anthropometric definitions for body measurement.

41.

Cheng

Kuzmichev

Adolphe

. A digital replica of male compression underwear. Text Res J. 2020; 90: 877–895.

42.

Chen X, Zeng X, Koehl L, et al. Optimization of human perception on virtual garments by modeling the relation between fabric properties and sensory descriptors using intelligent techniques. In: Laurent A, Strauss O, Bouchon-Meunier B, Yager RR (eds) Information processing and management of uncertainty in knowledge-based systems. IPMU. Communications in Computer and Information Science. Cham: Springer, 2014, pp.606–615.

43.

Lee

Sohn

Kim

. A study on representation of 3D virtual fabric simulation with drape image analysis II—focus on the comparison between real clothing and 3D virtual clothing. J Fashion Bus 2019; 15: 97–111.

A virtual e-bespoke men's shirt based on new body measurements and method of pattern drafting

Abstract

Keywords

Methods

Subjects

Acquisition of BMN

Application of BM N

Fit evaluation

Textile materials

Results

Pattern-oriented body measurements of neck and shoulder

Pattern-oriented body measurements of torso and arm

Customized shirt pattern block

Fit evaluation of e-bespoke shirts

Comparison between RtW, MtM, and e-bespoke shirts

e-Bespoke shirts in different styles

e-Bespoke shirts made of different fabrics

Discussion

Conclusion

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References

Acquisition of BM_N

Application of BM_N