resolving surface collisions through intersection contour minimization pascal volino nadia...
TRANSCRIPT
Resolving Surface Collisions through
Intersection Contour Minimization
Pascal Volino
Nadia Magnenat-Thalmann
MIRALab, University of Geneva
SIGGRAPH 2006
Abstract
Advanced methods are need for CD To recovery intersecting surfaces
Not rely on intersection regions More broader application Much simpler to implement
Outline
Introduction Related Work Description of the Method Results Conclusions
Introduction
Introduction
Early CD for cloth are restricted for cloth and volumes
Methods for ensuring adequate constrains are complex
Related Work
Related Work
Not only prevent but also repair the intersections
Common approach is to identify the “collision regions” by Volino et al.[1995] Retain a constant orientation with detection Impractical to implement
Related Work
Simpler approaches preventing surface intersections to occur or sprea
d Baraff et al [2003] use “flypapering” Bridson et al [2003] preserve wrinkle patterns Bridson et al [2002] the most comprehensive way
Related Work
A major contribution by Baraff et al [2003] Global Intersection Analysis method Track intersections by identifying closed contour Establish orientation correspondence Apply collision response method
Major restriction Necessity of closed
intersection contour
Related Work
A major contribution by Baraff et al [2003] Global Intersection Analysis method Track intersections by identifying closed contour Establish orientation correspondence Apply collision response method
Major restriction Necessity of closed
intersection contour
Goal
To overcome these limitations Intersection region Non-consist constrains Impractical to implement
Minimizing the length of the intersections contour Rather than spending time to identifying colliding
surface regions
Description of the Method
Minimizing Edge-Polygon Intersections
Integration with Collision response
The local and the Global Scheme
Description of the Method Resolution scheme
Define a collision scheme that induce a relative displacement
So as to reduce the
length of the
intersection contour disappearance
of the surface
intersection
Minimizing Edge-Polygon Intersections
Minimizing Edge-Polygon Intersections
The local and the Global Scheme
Sometimes not efficient if the contour is already quite
straight or the actions are the
largest on the region
Results
Some Test ExamplesPerformance and Limitations
Results
Integrated in a cloth simulation engine Based on particle systems for accurate
representation Broad-phase CD using AABB hierarchies Collisions are detected through edge-polygon
intersections
Some Test Examples
Untangling Cloth Surfaces Local scheme works poorly in this context Between 0.4 and 0.2 sec per iteration
Some Test Examples
Robustness Test Randomly ignoring 50% of all collisions detected Local scheme is fairly efficient in this context
Some Test Examples
The Ribbon Simulation 80000 polygons of
the ribbon, took
roughly 3 to 30 sec
per iteration Shear deformations
often break the
simulation
Some Test Examples
Garment Simulation Several layers of cloth Not always accurate
Performance and Limitations
The number of edge-polygon intersections remains very low Compared to the number of regular collisions
between mesh elements Resolve intersections progressively along the
simulation is more efficient
Performance and Limitations
In few configurations, the algorithm converges to a local minimum
Combined with intersection region identify
Conclusions
Conclusions
The method is general and not suffer from the limitations of existing method
Simple to implement The local scheme only relies on a simple
geometrical computation The global scheme efficiently solve very large
surface intersections easily be combined with most usual
approaches for CD
Question?