Neural Importance Sampling for Layered BSDFs
This project aims to improve the simulation of complex materials such as layered paint, fabrics and coated metals in computer graphics applications. We will develop a neural network-based sampling model that allows accurate and efficient light transport within these materials. In addition, we will integrate this model with general path-guiding techniques to minimize noise on both microscopic and macroscopic levels of light interaction. This collaboration strengthens ongoing research between ÉTS (Canada) and XLIM (France), and provides advanced training for students in modern rendering methods that are directly compatible with digital content creation pipelines in animation, visual effects and video games.
Voir la description complète du projetAdrien Gruson
Université de Poitiers
Computer science
Education
École de technologie supérieure
Globalink Research Award
