Cloth modeling and simulation algorithms have been a great challenge to researchers since the 1980s. The existing approaches can generally be classified into three categories: geometric, physical, and hybrid techniques [13]. Geometric methods [18] apply geometrical equations to create clothlike visualizations without considering the physical properties of the cloth. Physically based models [1-9, 12, 14-17, 20, 22] represent cloth as a collection of numerous tiny elements. The forces or energies associated with each element are calculated to simulate the behavior of cloth. The accuracy and efficiency of the simulation vary from method to method. The hybrid techniques [2,11,16,17,19] combine physical and geometrical methods and are used mostly to implement more complex simulation models.Physically based modeling methods include finite element models [1,4,5,8,15], finite volume models [9], and particle system models [3,6,7,14,22]. Among these models, the mass-spring model is a simple and powerful approach. It models the cloth with interactive particles and each particle is animated by integrating the force or the energy associated with it over time. Thus the shape of draping cloth is obtained.In general, two kinds of approach are used to increase the computer simulation speed. One is to employ a faster numerical integration method, and the other is to optimize the collision detection algorithm. In terms of the numerical integration method, Baraff et al. [2] have defined an implicit integration scheme that allows the system to take large steps in the simulation and to use hard constraints although the system is highly complicated. Meyer et al. [12] applied a hybrid explicit-implicit integration algorithm to perform a fast and stable simulation based on an elastic mass-spring model.
1Traditional collision detection algorithms are based on geometric object-space interference tests that are likely to be time consuming. Various computer software-based optimization techniques have been introduced to reduce the number of checks [1,11,15,17,19]. In addition, hardware-based techniques such as image-space tests have also been developed to speed up the process [16]. These approaches are very efficient, although their dependence on specific hardware may prevent their wide applications on the Internet.Abstract In order to perform virtual reality fitting of a garment through the Internet, an effective method of three-dimensional (3-D) garment modeling is developed. The garment is simulated using a 3-D quadrangular mesh based on a massspring system. To demonstrate the dynamical draping behavior of particular woven fabrics, the material properties adopted in the simulation, including tensile, shear, and bending, were measured using the Kawabata system. In animating the garment, a method of minimal enclosure was introduced to facilitate fast and reliable detection of cloth-body and cloth-cloth collisions.