Sampling hierarchical position-based dynamics simulation
Authors: Wang, M., Zheng, H., Qian, K., Li, S. and Yang, X.
Journal: Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume: 10582 LNCS
Pages: 45-55
eISSN: 1611-3349
ISBN: 9783319694863
ISSN: 0302-9743
DOI: 10.1007/978-3-319-69487-0_4
Abstract:The representation of detail is an essential part of animation. However, realistically and efficiently simulating details of folds and wrinkles in cloth has always been a huge challenge. Although the position-based dynamics method can simplify and generally depict details, grids are so fine that the simulation frame is far from satisfactory. The hierarchical position-based dynamics method provides an improved scheme. However, it is not capable of optimizing all grids effectively. In addition, during the coarsening process of the hierarchical selection procedure, some polygonal parts do not have effective convergence speed. We propose a voxelization-based sampling method. The proposed sampling method not only applies to any hierarchical grid but also avoids the uneven convergence speed of local simulation through particle selection. Experimental results show that the hierarchical sampling model proposed in this paper can accelerate the convergence of all layers of details.
Source: Scopus