Interactive Path Tracing and Reconstruction of Sparse Volumes
Nikolai Hofmann, Jon Hasselgren, Petrik Clarberg, Jacob Munkberg
Abstract
We combine state-of-the-art techniques into a system for high-quality, interactive rendering of participating media. We leverage unbiased volume path tracing with multiple scattering, temporally stable neural denoising and NanoVDB [Museth 2021], a fast, sparse voxel tree data structure for the GPU, to explore what performance and image quality can be obtained for rendering volumetric data. Additionally, we integrate neural adaptive sampling to significantly improve image quality at a fixed sample budget. Our system runs at interactive rates at 1920 × 1080 on a single GPU and produces high quality results for complex dynamic volumes.
Topics & Concepts
Path tracingComputer scienceRendering (computer graphics)VoxelLeverage (statistics)Volume renderingArtificial intelligenceRay tracing (physics)TracingComputer visionGlobal illuminationComputer graphics (images)PhysicsOperating systemQuantum mechanicsComputer Graphics and Visualization TechniquesAdvanced Vision and ImagingAdvanced Image Processing Techniques