VIRTUAL RECONSTRUCTION OF HISTORICAL MEN S SUIT Victor Kuzmichev 1, Aleksei Moskvin 2, Mariya Moskvina 2 1 Ivanovo State Polytechnic University, Textile Institute, 21, Sheremetev Av., Ivanovo, Russian Federation 2 Saint Petersburg State University of Industrial Technologies and Design, Department of Clothes Design and Technology, 18, Bolshaya Morskaya Str., Saint Petersburg, Russian Federation E-mail: wkd37@list.ru, lelikn2@mail.ru, petrushenkoma@yandex.ru Abstract: Keywords: Nowadays, the virtual technology is being widely applied in the area of clothing design and try-on. However, the possibilities of these technologies cover only the contemporary marketable clothes, while the insight in the aspect of historical costume is very limited. In this research, we developed the method that allows to reconstruct and do the virtual try-on of historical men s suit consisting from four different garments trousers, shirts, vest, and coat. The method includes, on one hand, the analysis of pattern drafting systems, patterns construction, special means of bespoke tailoring that were popular in the history and, on the other hand, the way of its adapting and preparing to contemporary technologies of 2D and 3D design. The exploration was done with men s suit and the patterns from the nineteenth century. We studied how the tailors took all measurements, the content of size charts including divisional, direct measurements, and its combination. To parameterize the historical patterns of men s clothes, we created the schedule of special indexes. We developed the method how to identify the means of garment shaping by steam pressing, which are hiding in the patterns, and how to perform ones by darts. The preparation of historical patterns to virtual try-on was done by CAD. As example, the reconstruction of full-dress suite painted on the Prince Albert of Saxe-Coburg and Gotha portrait (1840) was done, and high adequacy between the historical prototype and the virtual suit has been proved. Historical costume, reconstruction, VR, men suit, CAD, 3D 1. Introduction The reconstruction of historical costume is the new direction of the engineering design that can be applied in museums, theaters, and cinemas. Famous designers very often have taken the historical costume as an inspiration for new clothes designing. There are three objects saving the information about the historical costume such as real full-size clothes, printed images, and full-size and small-size patterns that were published in the illustrated books, the pattern-drafting books, and the pattern manuals. 2D or 3D objects mentioned have different content related to outline shape, construction, and style. There are many pictures, paintings, historical illustrations, engravings, and so on that have 2D images of historical costume in small, real, or big sizes as the resources of important information such as total view, silhouette, style, and different details. The pattern blocks that were created by measuring the garment or the body are the additional and technical-base supporting resource of bespoke tailoring, main measurements, ways of garment shaping, and so on. To do the real reconstruction of historical costume with high precision, we should extract the information from each object and consider all objects together. For example, the results of reconstruction will be different if we have only the image printed or when we use the image and the patterns published together. Virtual reconstruction is the second and parallel way of getting 3D images of historical costume. Technological platform for virtual reconstruction includes 3D methods of simulation of system avatar clothes and try-on. Many scientists are doing an exploration in this area. Martin and Ko [25] described the methods how to shape the 3D historical garments based on their pattern blocks and shown the possible areas of its application. Kang et al. [16-18] did the exploration about virtual and real reconstruction of historical costume. Magnenat-Thalmann developed the method of 3D modeling and simulating of historical garments in different styles in according with textile materials properties [24, 40]. Cybulska [10, 11], Crowfoot, Pritchard, and Staniland [9] studied the properties of textile materials used in the history and developed the process of virtual simulation. But obtaining the information and databases aren t enough for the reconstruction of historical clothes, either as real or as virtual objects. For example, to reconstruct 3D garments, Arnold [5, 6], Thursfield [33], and Waugh [37, 38] used the pattern blocks from the modern manuals without the consideration of its drafting algorithm. Aldrich explored the different aspects of the tailoring such as the changing of measurements used in the history, the system of sizing, the pattern-drafting methods of men s wear and http://www.autexrj.com 281
women s wear, the anthropometrical methods, the size tables, and the ways of pattern drafting and clothes shaping that the tailors and the dressmakers have used in the history [3,4]. Muller and Sohn [21-23] have systematized the huge information about the historical costume including patterndrafting algorithms and studied the methods of pattern shaping and sewing of historical garments by using the complex analysis of own real experience. So, to better understand the state of historical clothes shaping and to use this database in 3D simulation, it is necessary to know all aspects of male body morphology, pattern-drafting methods, pattern construction, and the relations in between. Contemporary mathematical-based methods of pattern construction can be put as the fundamental for accurate application of historical patterns in 3D CAD. This approach can help to avoid the subjective decisions during the virtual reconstruction and to get the realistic presentation of virtual garment. The goal of the current study is to develop the method of virtual reconstruction of historical men s suit into two categories, 2D and 3D. This method will use the information after complex analyzing of pattern-drafting manuals and its unification, the body measurements, and its application in pattern-drafting, sewing methods of pattern shaping. The main criteria of virtual reconstruction are the high reality of simulated system avatarclothes and huge adequacy between a real prototype and a virtual simulation. The flowchart of the developed method of virtual reconstruction of historical costume, which is based on the costume s image and adapted patterns, is shown in Figure 1. Figure 1 shows five steps that should be done to transform the historical image into the virtual system avatar-costume. The first step (named as 2D historical costume) is based on the 2D image of historical costume taken from the published books, journals, or paintings. The 2D image is the recourse from which the initial database could be built, such as the period of time when the costume was designed, the basic body measurements, the construction of garment, and the features of style to evaluate the costume fit from 2D image. Sometimes the chosen costume can be supported by the published pattern blocks. The second step preparation of initial database is devoted to the choosing and analyzing of pattern blocks by special indexes, body measurements and conditions of its taken from the body, and size charts used in chosen period of time. It s necessary because the chosen prototype has not published patterns. After this step, the full information should be completed about bespoke tailoring, system of sizing or size tables, applying of measurements to pattern drafting and construction, and system of pattern drafting (divisional system, direct measurements for pattern calculation, or systems which combined the both, indexes for accurate parameterization of pattern, etc.). The third step 2D reconstruction includes the adaptation of historical patterns to the similar positions of contemporary patterns positioned in rectangular net and the evaluation of the fabric deformations that are used to shape of the costume and its transformation during the manufacturing technology. The fourth step 3D reconstruction includes the designing of an avatar in accordance with the measurements used in the chosen period of time and virtual try-on simulation. The avatar should be an exact replica of historical prototype that has to have real-looking image, the same body measurements, and morphological features. When the virtual system avatar costume is designed, it should be compared with the historical prototype to check the adequacy. There are two possible situations that can take place after the comparison. If both objects the prototype and the virtual analog have the significant differences and not equal to each other, the pattern indexes, body dimensions, and properties of textile materials should be changed. The fifth step 3D virtual system avatar-costume includes the presentation of virtual system avatar-costume obtained with authentic and realistic view as the initial prototype. In our study, we investigated men s full-dress suite of the nineteenth century. This study includes a block patterns, styled patterns, 3D design and 2D pattern preparation, virtual garment prototyping, and prototype image. 1. 2D historical costume (prototype) 2. Preparation of initial database Choosing of historical pattern drafting system Choosing of size chart Pattern block indexes Body measures 3. 2D reconstruction Adapting of pattern block Generating of avatar Virtual try-on 4. 3D reconstruction Not equal Equal Comparison of 3D and prototype Adaptation the avatar poster to prototype 5. 3D virtual system "avatar-costume" Figure 1. Flowchart of virtual reconstruction of historical costume http://www.autexrj.com/ 282
2. Experimental 2.1. Men s suit The costume of Prince Albert of Saxe-Coburg and Gotha, 1840 (the painter is Sir William Ross) [28], was chosen for reconstruction. Figure 2(a) shows this costume titled as the full-dress suite and related to distinctively evening dress Correct dress [12]. There are four garments combined in this costume: the coat, the vest, the trousers made of high-quality black wool (worsted, either very narrow diagonal or fine wales), and the shirt made of linen or cotton with a linen or pique bosom. Figure 2(b) shows the total contour of the costume, Figure 2(c, d) shows only the visible parts of the vest, the shirt, and the trousers. Figure 2 demonstrated that the image of the costume contains the limited information of outline shape, silhouettes, and construction of the garments. The contour of the coat could be seen only in the front view, not on the back or side (Figure 2a), but the silhouette features of the vest (Figure 2b), the shirt (Figure 2c), and the trousers (Figure 2d) are completely hidden. So, the absence of information how all garments have been constructed needs more deep knowledge which we can find in pattern-drafting books published in present time. 2.2. Body sizes To build the Prince Albert s avatar, we should know the collecting body dimensions, methods of body measuring, and chart sizes used to draft the patterns. During the first half of the nineteenth century, the divisional systems that used the proportions of breast or a combination of the breast and the back length to calculate the coordinates of specific points were very widely spread. The number of measurements was limited and included only a few ones. For example, nine measurements were used to draw the coat pattern in the system The taylor s instructor, 1809 [27]. Full girths (bust, hip, etc.) and linear measurements were not divided between the front and the back, so the front width and the back width were calculated in proportions. Under the development of anthropometry and pattern drafting systems based on direct body measurements, more complex combination of body measurements such as arcs and diameters are used to the evaluation and presentation of male bodies. Figure 3 shows the schemes of bodies measuring in the nineteenth and twenty-first centuries. Figure 3(a) presents the collecting measurements from the drafting systems 1840, 1867, and 1891 [31, 1, 7]. Figure 3(b) shows the contemporary measurements from W. Aldrich s drafting system 2009 [2]. Contemporary methods of pattern making using the direct measurements in which points of the draft were identified with direct reference to body or garment measurements. The similar measurements belonging to historical and contemporary anthropometrical sizing systems are shown by solid lines; outdated measurements are shown by dotted lines. Table 1 contains the anthropometrical measures to be used in the nineteenth century and at present time (the numbers 1, 2, 3,... 32 are following Figure 3). As shown in Table 1, some historical measurements (5, 15 20, 22 26, 28, 29, 31) are not applied in modern anthropometrical programs. Special feature of drafting manuals of the nineteenth century is using the additional arc measurements for the presentation of upper bearing surface located between neck line, armhole line, and bust line. For example, there are an upper shoulder measurement (Figure 3, 19) and lower shoulder measurement (Figure 3, 20) that have been measured, respectively, across the armhole line from seventh cervical and the middle of back near the blade prominence. Front shoulder measurement (Figure 3, 15) is the circle from seventh cervical across the shoulder area to front armscye. The tailors in the nineteenth century also used the measurement Blade (Figure 3, 50) that is equal to half of the back width + armhole width. The sleeve length (9) was measured from the center of back as shown in Figure 3(a). To draft a trouser s pattern block, the heights 12, 13, 14, 29 and the arcs 2, 3, 29, 30, 31, 32 were measured as shown in Figure 3(a). (a) (b) (c) (d) Figure 2. The contours of Prince Albert of Saxe-Coburg and Gotha full-dress suite (a), and visible contours of the vest (b), the shirt (c), and the trousers (d). http://www.autexrj.com/ 283
(a) (b) Figure 3. Garment and body dimensions used in the nineteenth century (a) and W. Aldrich s system (b) [2] Table 1. Historical measures and its contemporary analogs No. Measure Measure No. Historical Contemporary Historical Contemporary 1 Breast Bust 17 Length at front - 2 Waist Waist 18 Length at the hip - 3 Seat Hips 19 Upper shoulder measure - 4 Back/half back Back width 20 Lower shoulder measure - 5 Blade - 21 Length to bottom of seat Body rise 6 Neck/around the neck Neck size 22 Back waist - 7 Back scye depth Armscye depth 23 Side length - 8 Length to natural waist Nape to waist 24 Back armseye depth - 9 Sleeve length Sleeve length + back width 25 Over shoulder - 10 Upper arm/biceps Top arm 26 Seye to waist - 11 Sleeve at cuff Wrist 27 Inseam length Waist to floor body rise 12 Length to seat line Waist to hip 28 Upper seat - 13 Length to knee Waist to knee 29 Thigh - 14 Length to sole of foot Waist to floor 30 Knee Knee 15 Strop / front shoulder - 31 Calf - 16 Over shoulder - 32 Bottom width Ankle The tailors in the first half of the nineteenth century took all measurements including the breast girth from their customer s garments such as coats, frock coats, and morning coats. These measurements are used to copy the ready-wear clothes. But this combination of such measurements couldn t help to parameterize the male body shape, to generate an avatar using computer graphics, or to test the fit of reconstructed historical costume. During the second half of the nineteenth century, the tailors took the measurements from customer s vest because the vest was close-fitting garment with smallest eases. As these measurements are more similar to the anatomical body measurements, we used ones from the size table published for generating an avatar. The schedule of historical measurements that were chosen for virtual simulation has been completed only by ones that have not the analogs in contemporary system of sizing and CAD modules. http://www.autexrj.com/ 284
For generating the avatar and drafting the pattern block, we applied the main arcs and heights to construct the shape and the proportions of body such as 1 14, 21, 27, 30, 32 published in the nineteenth century (Figure 3; Table 1). In addition, we measured some arcs on the upper surface of the body (15 20, 22 26, 28, 29, 31). All the mentioned measures were also used for drafting the pattern block. 2.3. Historical pattern blocks Before the reconstruction, we analyzed the published pattern manuals to enrich the initial information generated from a 2D image of the costume. Forty-eight pattern manuals of men s clothing published during 1809 1906 in Europe, the United States, and Russian Empire were investigated for extracting the following databases: - collection of anthropometrical data (method of measuring, sizing system construction, body measurements); - collection of pattern blocks (parameters, outline configurations, etc.); - algorithms of drafting or altering pattern blocks (such as the relations between direct or proportional measurements and the positions of patterns points and dimensions); - pattern-drafting methods; - means of garments shaping during its manufacturing. Figure 4 shows the row of pattern blocks and additional rectangular lines in manuals (1809, 1840, 1895, 2010). Figure 4(a) shows that the location of pattern has been changed from irregular that was similar to their position on the fabric before cutting to a regular position inside rectangular net. The second one was based on the overlapping the basic levels belonging to the body and the patterns levels (bust, waist, hip) in accordance with the grain direction. As we can see, different methods have been developed for using the different number of body measurements. Pattern manuals of the beginning of the nineteenth century were based on the body measurements that allowed to alter the pattern block in total and some parts, but the position of patterns was irregular because each pattern was corrected independently from others ones, see Figure 4(a). Under the development of mathematical-based pattern construction and the size tables of body measurements taken directly from the body, the tailors have drafted the patterns by using these measurements that allowed to find the coordinates of anthropometrical points in the reference to body measurements, see Figure 4(b). At the end of the nineteenth century, the patterns were built in rectangular net consisting main horizontal and vertical lines, see Figure 4(c). Figure 4(c, d) shows that the pattern position in 1895 was similar to contemporary basic pattern block in 2010. So, the variety of patterns and ways of its drafting in different manuals need the unique approach for its analysis. 3. Results and discussion 3.1. Parameterization of patterns The main aim of this step is to adapt the historical patterns to the possibilities of CAD modules and digital processing. For this purpose, we established the method how to parameterize the pattern blocks. Figure 5 shows the scheme how to measure the patterns of coat, frock coat, morning coat, vest, shirt, and trousers created in nineteenth century. We used 2D AutoCAD for measuring the patterns. To parameterize the patterns, we have developed the absolute and zero relative level indexes in four groups: 1. the lengths (a6, a7, b2, b3, c6, d1, R1, etc.). The indexes with a and b are related to the bodice and the collar (Figure 5 (a, b, d, e, f, g)), those with c are related to the sleeve (Figure 5 (b, e)), those with d and e are related to the trousers (Figure (a) (b) (c) (d) Figure 4. The pattern blocks: (a) frock coat with the skirt and front designed in one piece 1809 [27], (b) frock coat 1840 [30], (c) morning coat 1895 [14]), (d) frock coat designed in two piece 2010 [20]) http://www.autexrj.com/ 285
(a) (b) (c) (d) (e) (f) (g) Figure 5. Scheme of patterns measuring of men s clothing in nineteenth century. (a) Bodice of dress coat, frock coat, and morning coat; (b) sleevein and collar; (c) trousers; (d) bodice of shirt; (e) sleeve of shirt; (f) vest; (g) pockets. 5(c)), and those with R are related to the curved lines; 2. the angles between two lines (γ, β, δ1, δ2, etc.); 3. the ease allowance added to the body measurements or to the line of patterns such as a1, a2, a3, b1 (Figure 5a); 4. the proportions between the widths of two parts as /W1/ / W6/ (Figure 5a). The unit of indexes of first and third groups is millimeter, and that of second one is degree. The indexes of forth group have been calculated by dividing two values. To measure all indexes mentioned, we used the next procedure of pattern preparation. 1. The pattern blocks have been analyzed by means of the combination of specific parameters (eases, lengths, etc.). 2. Raster images of pattern blocks have been transformed into vector ones. 3. Patterns have been posed in the rectangular net. The patterns of bodice (coat and similar garments) have been joined along bust, waist, and hip levels, which were drawn as perpendicular to gain line (Figure 5a, d, f). The upper and down parts of sleeve were posed along the vertical gain line and the horizontal line across the lowest point of the cup of sleeve (Figure 5b). One-piece sleeve was posed along the vertical gain line (Figure 5e). The front and the back of trousers were joined along crotch, knee, and bottom in horizontal direction and along the folding line in vertical direction (Figure 5c). 4. The width and length of the rectangular net have been calculated. 5. Small-scale patterns have been transformed into full- http://www.autexrj.com/ 286
size patterns in accordance with the width and the length of rectangular net. 6. The indexes of first and second groups have been measured directly from the patterns. 7. The indexes of third group have been calculated by using the body measurements and the pattern parameters. 8. The indexes of forth group have been calculated by using the body measurements and rectangular net parameters. A total of 139 patterns, including 24 coats, 15 frock coat, 14 morning coats, 18 vests, 15 shirts, and 36 trousers, have been parameterized. The total number of measured and calculated pattern indexes were 54 for the coat, 53 for the frock coat, 53 for the morning coat, 26 for the vest, 28 for the shirt, and 37 for the trousers. The precision of parameterization was near ±1 mm because of 2D CAD, vectoring, scaling, and measuring. These indexes could be measured in parallel from the patterns or from the ready clothes; so by this approach, both the objects will be considered as one system for different purposes such as an identification, a classification, reconstruction, and creation of universal databases. The schedule of indexes is enough for the identification of the pattern construction of garments that formed the male costume during 1807 1906. 3.2. Parameterization of deformations along the front edge This step was devoted to how to produce the effects of steam pressing in virtual clothes and how to change this kind of deformation by darts along the front edge. Steam pressing or heat-moisture treatment was widely used for the deformation of clothes in manufacturing process in the nineteenth century [8; 26; 29; 32]. The aim of this process is to adapt the shapes of clothes to human bodies or fashionable style. Figure 6 shows the areas of historical patterns that must be treated by steam press for shaping and fitting the body. Figure 6 shows the places of steam pressing in all dressmaker manuals. For coats, after steam pressing, the length of edges should be elongated or shorten: in area 1, from 0 to 9 mm; in area 2, from 6 to 13 mm; in area 3, from 6 to 64 mm; and in area 4, from 6 to 25 mm. For trousers, the dressmaker s manuals have recommended the next deformations: in area 5, from 0 to 9 mm; in area 6, from 12 to 29 mm; in area 7, from 23 to 31 mm; in area 8, from 32 to 59 mm; and in area 9, from 19 to 37 mm. The edges in areas 5, 7, and 9 should be elongated under steam pressing, but in other areas 1 4, 6, 8, these should be shorten. But contemporary technologies of men s wear production don t need such huge deformations. So for accurate reconstruction of chosen costume, we should consider all means of shaping hidden in the pattern blocks. The values of deformations can be calculated as the difference between the lengths of sewing edges belonging to the front and the back. But to recognize the deformation of front edge to change its configuration from convex to straight is impossible. To find the relations between the front edge configuration and the values of its deformation, we studied many pattern blocks designed with steam pressing or without steam pressing [13]. Figure 7(a) shows the construction of initial right front consisting the side and the central parts that we used to establish new method. The front shaping is based on two approaches: first, by the vertical waist dart located between the side part and the central part and, second, by decreasing the neck line. (a) Figure 6. The places of steam pressing: (a) coat [13, 8] and (b) trousers [34] (b) http://www.autexrj.com/ 287
(а) (b) (c) Figure 7. Scheme of front side part adapting: (a) pattern of initial parts (1906) [12], (b) pattern shaping by means of five darts from neck and front edges, and (c) position of five darts We have shaped the front parts as shown in Figure 7(b). 1. We calculated the dart length (in cm) by using the following equations: lb = 0.4 (FW + ease), (1) lw = 0.35 (FW + ease) (2) where lb is the length of dart located on the bust level, lw is the length of dart located on the waist level, and FW is the front width. 2. We measured the distance L between bust and waist levels (Figure 7c). 3. We divided the front central part into five segments A1 A5 in 0.3L (thick dotted line, Figure 7b). 4. We have drawn the central lines of darts as the perpendiculars to the edge of central part. 5) We have drawn the neck dart instead of steam pressing deformation on the front side part. 6) We moved the segments A1 A5 from the front central part to side one into new positions В1 В5 (thick solid line, Figure 7b). After opening the new darts on front side part, the vertical waist dart should be closed. Figure 7(c) shows the modified front with four darts along front edge instead of the deformation after steam pressing. 3.3. Preparation of historical patterns for virtual try-on To do a virtual try-on and to adapt the historical pattern to contemporary modules of CAD, we used the relationship between the body measurements and the patterns parameters. After comparison of all the methods of pattern drafting that existed in nineteenth century, we presented the common rules of rectangular net drawing in Figure 8. The width of rectangular net (Figure 8) is equal to the sum of half bust girth and an ease allowance, the length is equal to the garment length. The bust, waist, and hip positions are marked by means of the sums of the measurements such as back scye depth, waist height, and seat height (or length to seat line ) and an ease allowance. The position of horizontal line across the lowest point of armhole is in an ease value down than the bust line. The waist darts WD and the values FBO of the front and the back overlapping on hip level have been calculated as the differences by using the following equations WD = [(BG + ease) (WG + ease)]/2, (3) FBO = [(BG + ease) (HG + ease)]/2, (4) where WD is the sum of waist darts designed along waist level, FBO is the value of the front and the back overlapping along hip level, and BG, WG, and HG are the girths of bust, waist, and hip, respectively. These principles of adapting and describing historical patterns, including the rectangular net drawing, the position of patterns inside the rectangular net, and equations for the calculation of darts and overlapping, are important to ensure the relations with contemporary methods of pattern drafting. To draw the patterns, we used the body measurements from chart sizes published in The American coat, vest and trousers system [14] because this manual has the biggest number of measurements. The Prince Albert s body was perfect in accordance with beauty rules in the first half of the nineteenth century...beautiful figure, broad in the shoulders, and a fine waist [39]. We have corrected some measurements and used the values: bust girth is 37 inch (94 cm), height is 68 inch (172.7 cm), half back is 7 inch (17.8 cm), blade is 11 inch (27.9 cm), waist girth is 28 inch (71.1 cm), seat girth is 38 inch (96.5 cm), thigh girth is 22 inch (55.9 cm), knee girth is15.5 inch (39.4 cm), back scye depth is 8.5 inch (21.6 cm), back length to natural waist is 17 inch (43.2 cm), length to seat line is 24.5 inch (62.2 cm), length to bottom of seat is 27 inch (68.6 cm), length to knee is 41 inch (104.1 cm), length to sole of foot is 59.5 inch (151.1 cm), and sleeve length is 31.5 inch (80 cm). http://www.autexrj.com/ 288
Figure 8. The adaptation of historical patterns of coat to contemporary rules of pattern drafting. The systems of pattern drafting and some constructive parameters were taken from The tailor s master-piece (1840) [31] and A treatise on cutting garments to fit the human form (1841) [30]. From tailor manuals, we took the recommendations how to shape the patterns under next sewing and heatmoisture treatment (or steam press). We transformed these recommendations in accordance with the previous method by the calculation of the darts, which should be design instead of the deformations. For trousers, the location of deformations was chosen in accordance with the properties of textile fabrics and fashionable silhouette. We used AutoCAD for pattern drafting. Figure 9 shows the pattern blocks of all garments from which the full-dress suit was combined. Dotted lines show the positions and values of darts that can be designed instead of heat-moisture treatment. After these modifications, all historical patterns were similar to the rectangular pattern construction that could be adapted into contemporary CAD. 3.4. Virtual try-on The virtual try-on of historical costume was made by software Clo3D. This software and other popular ones such as Marvelous Designer, Optitex, 3ds Max, DC Suite, and Vidya used the similar algorithm. Before the virtual simulation, we have chosen the same rare fabrics as in nineteenth century to be sure to copy the silhouette of the prototype. The next fabrics with specific physical and mechanical properties were used as presets in garment simulations: cotton for shirt, cotton pique for the front of vest, silk satin for the back of vest, wool for coat and trousers. Also we took into consideration the position of supportive trims that were mainly applied between two layers of fabric or on back side of fabrics for maintaining the actual shape and comfort of garments. The supporting fabrics made of cotton or linen were an essential accessory in garments, which we used in collar, cuffs, shoulder, and the front part of coat. The location of these supporting fabrics has been chosen in accordance with the tailors manuals [15, 19, 35, 36]. For the front part and the collar of coat, the front part of vest, the additional front part of shirt, and the cuffs of shirt, the physical and mechanical properties such as bending, buckling ratio, buckling stiffness, internal damping, and density were changed. Textures of fabrics were designed under consideration of colors and structures on the prototype. Under virtual try-on, we coped the order of Prince Albert s dressing by chosen garments (shirt-trousers-vest-coat) and modeled the interactions between the garments. Figure 10 shows the virtual systems avatar-full-dress suit 1840 formed on different steps of dressing. http://www.autexrj.com/ 289
(a) (b) (c) Figure 9. Parameterized and adapted patterns (1840) with the positions of deformations obtained under steam pressing (heat-moisture treatment): (a) coat, (b) trousers, (c) shirt, (d) vest (d) 3.5. Evaluation of reconstructed 3D systems To evaluate the precision of our method, we compared three contours of next objects: 1. 3D system avatar-full-dress suite 1840 generated from the patterns with the simulation of deformations under steam pressing (moisture-heat treatment); 2. 3D system avatar-full-dress suite 1840 generated from the patterns without taking into consideration the deformations under steam pressing; 3. the prototype. We put all images in the same scale and the same poses as Prince Albert standing. Figure 11 shows the contour of the full- http://www.autexrj.com/ 290
(a) (b) (c) Figure 10. Virtual systems completed layer-by-layer: (a) avatar + shirt + trousers, (b) avatar + shirt + trousers + vest, and (c) avatar + full-dress suite http://www.autexrj.com/ 291
(а) (b) (c) Figure 11. Contours of suits: (a) with steam press simulation, (b) without steam press simulation, (c) overlapping three contours and the measurements location Figure 12. The distances between the contours of virtual and real initial suits dress suite that combined all the mentioned garments objects. The next horizontal and vertical distances between the contours have been measured: Х1 the distance between the lapel edges near the neck, Х2 the distance between the lapel edges on the waist level, Х3 the distance between the front edges near the bottom, Х4 the distance between the bottom positions, Х5 the distance between the side contours on waist level, Х6 the distance between the side contours on hip level. Figure12 shows the measured values in millimeters. Figure 12 shows that the virtual coat reconstructed by ignoring steam pressing (without moisture-heat treatment) along the front edge has the biggest errors of edge length and wrong http://www.autexrj.com/ 292
configurations of front edge, lapel folding and side contour (dash-and-dot line): the values are X1 = 12 mm, Х2 = 42 mm, and Х3 = 16 mm. These values indicated how the ignorance of the deformations during the steam pressing can affect the silhouette. The value of X4 = 24 mm shows that the length of front edge is bigger than the prototype has in reality. The folds and the misfit in hip area have happen under the ignorance of skirt deformations; the value X6 = 57 mm indicated this influence. As shown in Figure 12, the reconstructed coat with the simulation of steam pressing (with moisture-heat treatment) has coped the outline shape of historical prototype with great precision. The differences between the initial and virtual garments are near to 8 mm for Х1 Х4, X6 and 12 mm for X4. The reasons of small deviations are appearing from the average body measurements that were used to the reconstruct the Prince Albert s body and the similarity of initial and reconstructed postures. 4. CONCLUSIONS The method of virtual reconstruction of historical clothes based on the 2D images initial picture of clothes and patterns used have been developed. The pattern manuals of men costume and the charts of body measurements in the nineteenth century were analyzed. To describe the lengths, the line configurations and other parameters of pattern blocks and its relationship with the body measurements, the schedule of special indexes has been created. New graphic method that allows to calculate and to identify the hidden elongation and the shrinkage of textile materials in historical patterns was promoted. Graphic schemes of the patterns parameterization of garments forming the full-dress suite 1840 such as coat, frock coat, morning coat, vest, shirt, and trousers were created. The schedule of indexes including the body measurements, the eases, and the parameters of pattern block can be applied as initial database in CAD for designing the garments with historical features. We simulated the system avatar-full-dress suite 1840 by covering and dressing of avatar layer by layer by different garments to get completely realistic view and the adequacy to historical prototype. The possible applications of developed method are different areas such as museum virtual exhibitions and performances; scientific classification considering two objects, 2D patterns and 3D garments; virtual reconstruction for organizing the virtual exhibitions; and contemporary design based on the historical prototypes. The application of developed method will allow to reconstruct the historical costume with the greatest precision and special features in 3D by using the historical patterns and published images. This method can help to save the lost information about the construction of historical costume and to enrich the professional treasure. ACKNOWLEDGMENTS This work has been fully supported by the Ministry of Science and Education of Russian Federation under the project 2.2425.2017/ПЧ Development of software for virtual design of system body clothes in static and dynamic and for virtual try-on FashionNet. References [1] Acton, WM.R. (1867). Acton`s improved system of actual measurement. William R. Acton & Co. (New York). [2] Aldrich, W. (2006). Metric Pattern Cutting for Menswear. Blackwell Publishing Ltd. (Oxford). [3] Aldrich, W. (2002). Pattern Cutting for Women s Tailored Jackets: Classic and Contemporary. John Wiley and Sons (Oxford). [4] Aldrich, W. (2007). History of sizing systems and ready-towear garments. In Ashdown, S.P. (ed.) Sizing in clothing: Developing effective sizing systems for ready-to-wear clothing. Woodhead Publishing (Cambridge). pp. 1-56. [5] Arnold, J. (1977). Patterns of fashion: Englishwomen s dresses and their construction. Macmillan (London). [6] Arnold, J. (1985). Patterns of fashion: The cut and construction of clothes for men and women. 1560-1620. Drama publishers (Los Angeles). [7] Arvidson, P.N. (1891). The Arvidson coat and vest system. Published by the author (Kalamazoo). [8] Berris, K. (1906). School of men s and ladies dress cutting (in Russian). Published by the author (Mitawa). [9] Crowfoot, E., Pritchard, F., Staniland, K. (2001). Textiles and clothing 1150-1450. The Boydell Press (Woodbridge). [10] Cybulska, M. (2012). To see the unseen - Computer graphics in visualisation and reconstruction of archaeological and historical textiles. In Mukai, N. (ed.) Computer Graphics. InTech (Rijeka). pp. 213-228. [11] Cybulska, M., Maik, J. (2007). Archaeological textiles - a need of new methods of analysis and reconstruction. FIBRES & TEXTILES in Eastern Europe, Vol. 18, No. 3 (80), pp. 100-105 [12] De Berard, F.B. (1887). Correct dress. J.W. Gibson, publisher (New York). [13] First auxiliary society of Saint-Petersburg cutters. (1906) Full Academic class of men s dress cutting of the First Auxiliary Society of Saint-Petersburg cutters (in Russian). R.Golike printing house (Saint Petersburg). [14] Gordon, S.S. (1895). The American coat, vest and trousers system. J.J. Mitchell co. (New York). [15] Jones, A. (1822). Rules and directions for cutting men s clothes. J. W. Copeland (Middlebury). [16] Kang, Y., Wu, S., Ko, Y., Kang, Z.Y., Kim, A., Kim, N. and Ko, H. (2013). Digital production of traditional costumes. In: Proceedings of the 1st International Conference on digital fashion. London: Digital Fashion 2013 Secretariat, 16-17 May 2013. London, pp. 329-335. http://www.autexrj.com/ 293
[17] Kang, Z.Y., Cassidy T.D., Li, D. and Cassidy, T. (2014). Historical Costume Simulation. In: 2014 International Textiles and Costume Culture Congress, 24-25 October 2014, National Chonbuk University, S. Korea. Available from: http://eprints.whiterose.ac.uk [Accessed 15/05/17]. [18] Kang, Z.Y., Young, C., Tracy, D., Cassidy, T. and Li, D. (2015). Historic costume simulation and its application. In: Proceedings of the 15th Autex World Textile Conference, 10-12 June 2015, Bucharest, Romania. Available from: http://eprints.whiterose.ac.uk [Accessed 14/10/16]. [19] Kennedy, H.G. (1891). The science of coat and vest cutting. Published by the author (Chicago). [20] M. Muller & Sohn. (2010). HAKA Schnittkonstruktionen Sakkos (in german). Rundschau-Verlag Otti G. Koniger & Co. (Munchen). ISBN.: 978-3-929305-57-9 [21] M.Muller & Sohn. (2010). Historische Schnitte: Schnitte für historische Gewänder (in german). Rundschau-Verlag Otti G. Koniger & Co. (Munchen). ISBN.: 978-3-929305-55-5 [22] M.Muller & Sohn. (2011). Schnittkonstruktionen Trachtenmode für Männer (in german). Rundschau-Verlag Otti G. Koniger & Co. (Munchen). ISBN.: 978-3-929305-59-3 [23] M.Muller & Sohn. (2014). Historische Schnitte: Grund- und Modellschnitte von historischen Kostümen. (in German). Rundschau-Verlag Otti G. Koniger & Co. (Munchen). ISBN.: 978-3-929305-72-2 [24] Magnenat-Thalmann, N., Volino, P. (2005). From early draping to haute couture models: 20 years of research. The Visual Computer, vol. 21, pp. 506-519. [25] Martin, K., Ko, H. (2009). Virtual historic costume across cultures and disciplines. In: Proceedings of the 15th International Conference on Virtual Systems and Multimedia, 9-12 September 2009, Vienna, Austria, pp. 104-108. [26] Matheson, H.H. (1871). Matheson s scientific and practical guide for the tailor s cutting department. Published by the author (New York). [27] Queen, J., Lapsley, W. (1809). The taylor s instructor, or a comprehensive analysis of the elements of cutting garments of every kind. Published by the author (Philadelphia). [28] Royal Collection Trust. (2017). Sir William Ross: Prince Albert, 1840. Retrieved June 11, 2017. Web site: https:// www.royalcollection.org.uk. [29] Schweinhart, W. (1851). The young crook`s guide, or selfvarying system for cutting garments. N.C. Neinstedt (York Sulphur Springs). [30] Scott, A. (1841). A treatise on cutting garments to fit the human form. Published by the author (New York). [31] Scott, A., Wilson, J. (1840). The tailor s master-piece. Published by the author (New York). [32] Stone, C.J. (1887). Stone`s paramount cutter: a system for cutting garments. Mever & brothers (Washington, D.C.). [33] Thursfield, S. (2001). The medieval tailors assistant: making common garments 1200-1500. Ruth bean Publishers (Bedford). [34] Vincent, W.D.F. (1898). The cutter`s practical guide to cutting and making all kinds of trousers, breeches and knickers. The J. Williamson company limited (London). [35] Vincent, W.D.F. (1898). The cutter`s practical guide to cutting and making all kinds of waistcoats. The J. Williamson company limited (London). [36] Vincent, W.D.F. (1898). The cutter`s practical guide to cutting, making and fitting lounges, reefers, norfolk, sporting, and patrol jackets. The J. Williamson company limited (London). [37] Waugh, N. (1964). The cut of men`s clothes, 1600-1900. Theatre Arts Books (New York). [38] Waugh, N., Woodward, M. (1968). The cut of women`s clothes, 1600-1930. Theatre Arts Books (New York). [39] Williams E.T., Veldman, M. (2017). Victoria, Queen of United Kingdom. Retrieved June 17, 2017. Web site: https://www.britannica.com. [40] Yasseen, Z., Nasri, A. H., Boukaram, W., Volino, P., Magnenat-Thalmann, N. (2013). Sketch-based Garment Design with Quad Meshes. Computer-Aided Design, vol. 45, Issue 2, pp. 562-567. http://www.autexrj.com/ 294