System and method for surfacing of particle systems
View Patent ↗A method and system for surfacing of particle systems is proposed to achieve high surface resolutions and low execution times, comprising partitioning particles into sub-blocks, each sub-block associated linear data arrays, sequentially accessing each array to composite the particles into a respective sub-volume using estimates of grid stencils derived from particle kernels, performing random-access into the sub-volumes, processing each linear array, and lexicographically pushing the corresponding sub-volume into a DT-Grid to convert the sub-volume to a narrow band level set. Further steps to improve the temporal coherence of resulting surface, to reduce artifacts from sparsely sampled regions of space, to enhance artistic expression, and for flexible post processing can be performed.
1. A method for spatial anti-aliasing in a particle system, comprising:
computing a physical simulation with a particle system, wherein the particle system includes a plurality of particles, and wherein each of the plurality of particles retains a static shape during computation of the physical simulation;
subsequent to completion of the physical simulation, replacing particles within the particle system with dynamic shapes, wherein each dynamic shape is associated with a plurality of shape variables;
calculating each shape variable for each dynamic shape based on its position relative to neighbors and a neighbor density;
diffusing the dynamic shapes in accordance to the calculated shape variables, wherein the diffused dynamic shapes form a diffused surface; and
outputting the diffused surface.
2. The method of claim 1 , wherein the shape variables for each dynamic shape is calculated with tensor algebra.
3. The method of claim 1 , wherein each dynamic shape is an ellipsoid.
4. The method of claim 1 , wherein the physical simulation is computed with a smoothed particle hydrodynamics system.
5. The method of claim 1 , further comprising:
modifying the diffused surface with noise functions and scaling.
6. The method of claim 1 , wherein the particle system is stored in a level set data structure residing in an accessible memory.
7. The method of claim 6 , further comprising:
performing morphological operations on the level set data structure to remove holes and disconnected pieces in the diffused surface.
8. The method of claim 1 , further comprising:
performing a surface smoothing on the level set data structure with a mean-curvature based level set flow.
9. The method of claim 1 , wherein particles in the particle systems are spheres.
10. The method of claim 1 , wherein particles in the particle systems vary over time.
11. A method for spatial anti-aliasing in a particle system, comprising:
computing a physical simulation with a particle system, wherein the particle system includes a plurality of particles, and wherein each of the plurality of particles retains a static shape during computation of the physical simulation; and
outputting the physical simulation as a diffused surface, wherein:
particles within the particle system are replaced with dynamic shapes, wherein each dynamic shape is associated with a plurality of shape variables;
each shape variable for each dynamic shape is calculated based on each dynamic shape's position relative to neighbors and a neighbor density; and
the dynamic shapes are diffused in accordance to the calculated shape variables, wherein the diffused dynamic shapes form a diffused surface.
12. The method of claim 11 , wherein the computation of physical simulation involves generating a plurality of simulation frames.
13. A method for spatial anti-aliasing in a particle system, comprising:
computing a plurality of simulation frames with a particle system, wherein the particle system includes a plurality of particles, each of the plurality of particles representing a spatial area in a given simulation frame and wherein each of the plurality of particles retains a static shape during computation of each of the plurality of simulation frames; and
outputting each of the plurality of simulation frames previously computed as a diffused surface, wherein:
particles in a given simulation frame are replaced with dynamic shapes, wherein each dynamic shape is associated with a plurality of shape variables;
each shape variable for each dynamic shape is calculated based on each dynamic shape's position relative to neighbors and a neighbor density; and
the dynamic shapes are diffused in accordance to the calculated shape variables, wherein the diffused dynamic shapes form a diffused surface.