What-you-see-is-what-you-get Friction for computer graphics
Surface roughness is a well-studied material property in the context of computer graphics and sophisticated approaches for modeling the physical behaviour of reflectance due to roughness have been developed (e.g., microfacet BRDFs). On the other hand, methods for modeling interaction between surface in multibody physics simulations often lack the ability to capture rich details due to roughness. Particularly, Coulomb friction is popular due to the fact that it is easy to compute and integrate with existing simulation pipelines, but is ultimately a coarse approximation.
Furthermore, despite roughness being an important aspect of both reflectance and friction, there is currently no correspondence between the models used for both simulations.
This project will investigate unifying the surface roughness models used in computer and aims to develop a more holistic approach for integrating roughness in computer graphics applications. Specifically, the goals of the project will be to:
- Capture the frictional and reflectance behaviour of real-world surfaces due to surface roughness;
- Use machine learning techniques to learn frictional models parameterized by their surface roughness;
- Learn correspondences between these parameterized friction models and the reflectance models used in physically based rendering.
Experience with computer graphics and multibody simulations and physics simulation. C++ and MATLAB experience also recommended. Fluent in French or English.
Programme d'études visé
Maîtrise avec mémoire, Doctorat
Domaines de recherche
Technologies de l'information et des communications
Une bourse de professeur / funding is available
Date de début : 2020-01-01