【正文】 description of how the garment is mapped from 3D to 2D which will enable the reverse process to be achieved when draping is authors have described elsewhere [4], a flattening algorithm which operates on a polygon list of 3D triangles to provide 2D flattening. This algorithm is capable of handling the arbitrary siting of seams which could involve darts or gussets depending on the nature of the curvature involved. . Drape engine This module should be capable of taking the following information: ● A geometric description of a 2D pattern (including sufficient interior points and connectivity relationships with other pattern pieces). ● Fabric type as described by key material characteristics. ● Constraining mechanisms, . straps, zips. ● Underlying mannequin surface description. ● Surface texture description. and produce an accurate prediction of the final shape of the fabric. This is avery difficult and putationally demanding approaches are possible as summarised elsewhere [5]. The one taken by the authors has been to model the various energy dissipation modes which exist when a garment drapes. These result from tensile stiffness, bending stiffness and buckling behaviour. The precise variables used to represent material behaviour are tensile strain energy constant,inthewarpdirection,Ksu。 potential energy resulting from fabric mass,Kg. The model developed [6], evaluates the total energy derived from these sources and obviously depends on the particular fabric used. When individual nodes are then moved in 3D to reduce the total energy, major problems can arise. Checks must be continuously made to ensure that nodes on the garment panel do not intersect with the underlying surface of the mannequin. This can be handled by incorporating an additional energy ponent to the list above, which severely penalises nodal movement if intersecting the mannequin surface. Additionally, there is the problem of the 3D panel actually intersecting with factors serve to dramatically increase the putational plexity of the problem. The model, which embodies these energy and geometric modelling elements, is termed a drape engine here. The method of presentation and how close it is to the final shape assumed by the pattern is important. Previous work in this area has highlighted the putationally demanding nature of solutions and the sensitivity these solutions have with respect to the initial 3D position. 4. Example Suppose that a designer is required to design the right front panel of a fitted will require the idealised specification of the 3D surface for the panel and the optional location of where a dart can be sited to improve the fit of the panel. Two possible examples are presented. Since the examples are not part of a selfsustaining garment, then fixture points will have to be specified or the panels will simply slip off the mannequin when submitted to the drape engine. Fig. 3(a) illustrates the initial representation of the first garment panel indicating the idealised 3D shape, the positions of fixation to the underlying body (points A, B, Cand D). No dart is specified for this example. At this stage,the surface is represented by a polygonal mesh. The precise nature of the mesh is shown in (b).Two fabric types A and B, will be considered for the two examples and the material characteristic sare defined in Type A is nominated for the first flattening process is now carried out to obtain the 2D pattern. This provides a onetoone mapping for each of the nodes in the initial 3D mesh(Fig. 3(b)) in proceeding from 3D to 2D. The result is provided in Fig. 3(c). Finally, the drape engine is starts with the initial 3D shape of Fig. 3(a) and seeks to minimise the total energy required to distort the 2D flattening to assume the current 3D shape. The final 3D draped shape is provided in Fig. 3(d) which takes into account the fixing positions for the panel at points A, B, C and D. For added visual realism, texture rendering is applied. This is enhanced because of the authentic 2D– 3D mapping which is provided by the drape engine. A second example is provided in (a)– (d).In this case another fabric (Type B) is used and a dart is specified in the initial 3D specification of the panel (Fig. 4(a)). The initial 3D surface is triangulated in such a way as to preserve the integrity of the dart geometry (Fig. 4(b)). This allows the flattening to proceed (Fig. 4(c)). Nodes along the dart line are duplicated which causes a dart to be formed during the flattening process. During the progress of the drape engine, dart nodes are forced to be located at the same 3D final result is depicted in Fig. 4(d).The two examples highlight the contribution which fabric properties and tailoring mechanisms such as darts have on the actual appearance of both examples originate from the same initial 3D surface, their final appearances can be dramatically different. 5. Conclusion Outlined in this paper is a framework for true integrated manufacture being applied to garment manufacture. The problems of the design interface, flattening and drape have been highlighted as being the main causes of lack of progress. The examples described, illustrate how final shape depends on fabric used and the incorporation of tailoring mechanisms such as darts. In order for a CAD system to be of practical use to garment designers and manufacturers it must facilitate these features in predicting the appearance of real life garments