| dc.description.abstract | Fluids such as gases and liquids are often animated through physically based techniques rather than manually. While this method can generate highly detailed fluids without tedious work from an artist, it also diminishes the amount of control an artist has over the end simulation. One proposed solution to this problem is to provide a low resolution preview to allow an artist to visualize a potential simulation before committing to a final, high resolution animation. However, increasing the resolution will also result in non-negligible differences between the preview and final versions of a fluid animation. A potential technique to combat this problem is match point sampling. In this scheme, match points are placed in regions of interest and are used as a guideline to force the fluid in that region to conform to how it behaved during its preview stage. This method has shown promising results, but its development has been limited to grid-based Eulerian simulations, which are primarily utilized for gas simulations. This work focuses on adapting the original sampling technique for use in particle-based Lagrangian simulations, which are primarily used for liquid simulations. Specifically, we investigate different methods of distributing and moving match points through three-dimensional space in the absence of a fixed grid and evaluate their efficacy by examining any differences between the preview and final versions of different liquid simulations. | |
