Abstract
Realistic garment simulations are required not only for academic research but also for use in practical applications such as computer-aided garment design systems. There have been several previous approaches to simulate garment deformation, including the well-known continuum mechanics and fsinite element method. Although, the particle-based method has been more favored in the area of computer graphics, because of its numerical stability and calculation speed. However, the calculation of bending force and its derivatives is a major time bottleneck, along with the collision detection and linear system solving process. Therefore, this study proposes a new theoretical framework with a vector-inner-product-based bending force term to avoid heavy calculations incurred through the conventional angle-based formula. The commercial software $Mathematica^{(R)}$, was adopted to derive the tensor products efficiently; in addition, the final calculation speed benchmarking was tested using a three-pieced female garment data set. The sparsity of the final linear matrix was decreased and the resulting final calculation speed-up was 3.3%, for a virtual female garment dress simulation with 7,783 triangular elements.