Flow-based method
Comparison of random seeding, regular seeding, Poisson sphere seeding,
flow-based template seeding and
template seeding with Poisson seeding.
|
|
|
| Random Seeding | Regular Seeding | Poisson Seeding |
|
|
| Template Seeding | Template plus Poisson Seeding |
|
|
| Different types of 3D critical points:
a,b) spiral c,d) spiral saddle, e,f) saddle g,h) source, sink i,j) center |
Seeding templates for various critical points.
The dots represent the seed locations in the templates for: a) spirals,spiral saddles, centers, b) saddles, c) sources and sinks |
Feature-based filtering
Feature-based streamline filtering uses multi-pass filtering
according to different geometric criteria.
|
|
|
| Before any filtering, 3168 streamlines. | After removing short streamlines, 2687 streamlines. |
After removing similar streamlines, 72 streamlines. |
|
|
| The 2615 similar streamlines that were removed. |
After removing streamlines with tight winding angles, 54 streamlines. |
Hybrid method
Perform template seeding with dense Poisson seeds
to fill empty space.
Apply filtering approach after that.
Same data set as previous examples:
|
|
| Template plus Poisson seeding
with 396 streamlines |
After filtering down to 49 streamlines |
Dataset with 108 critical points:
|
|
|
| Poisson Seeding with 403 streamlines |
Template seeding,before filtering with 5862 streamlines |
Template seeding,combined filtering down to 440 streamlines |
Hemisphere CFD dataset
|
|
|
| Random seeding with 99 streamlines |
Poisson seeding with 99 streamlines |
Regular seeding with 96 streamlines |
|
|
|
| Template seeding, before filtering with 107 streamlines |
Template seeding, after filtering with 72 streamlines |
Template plus Poisson Seeding with 99 streamlines |
Post CFD dataset(Template seeding with 70 streamlines)
|
|
| Top view | Front View |
Return to AVIS home page.