Real-time neural radiance caching for path tracing

  title={Real-time neural radiance caching for path tracing},
  author={Thomas M{\"u}ller and Fabrice Rousselle and Jan Nov'ak and Alexander Keller},
  journal={ACM Transactions on Graphics (TOG)},
  pages={1 - 16}
We present a real-time neural radiance caching method for path-traced global illumination. Our system is designed to handle fully dynamic scenes, and makes no assumptions about the lighting, geometry, and materials. The data-driven nature of our approach sidesteps many difficulties of caching algorithms, such as locating, interpolating, and updating cache points. Since pretraining neural networks to handle novel, dynamic scenes is a formidable generalization challenge, we do away with… 
Dynamic Diffuse Global Illumination Resampling
This work proposes a new real-time solution called dynamic diffuse global illumination resampling that computes direct and indirect illumination accurately and with low noise and outperforms hardware accelerated path tracing in both runtime and noise.
Neural radiosity
This work introduces Neural Radiosity, an algorithm to solve the rendering equation by minimizing the norm of its residual, similar as in classical radiosity techniques, and proposes a network architecture using geometric learnable features that improves convergence of the solver compared to previous techniques.
Instant Neural Graphics Primitives with a Multiresolution Hash Encoding
Fig. 1. We demonstrate instant training of neural graphics primitives on a single GPU for multiple tasks. In Gigapixel image we represent a gigapixel image by a neural network. SDF learns a signed
Dynamic Diffuse Global Illumination Resampling
Interactive global illumination remains a challenge in radiometricallyand geometrically-complex scenes. Specialized sampling strategies are effective for specular and near-specular transport because


Spatiotemporal reservoir resampling for real-time ray tracing with dynamic direct lighting
This work introduces a new algorithm---ReSTIR--- that renders such lighting interactively, at high quality, and without needing to maintain complex data structures, and derives an unbiased Monte Carlo estimator for this approach, and shows that it achieves equal-error 6×-60× faster than state-of-the-art methods.
Practical Path Guiding for Efficient Light‐Transport Simulation
This work proposes an adaptive spatio‐directional hybrid data structure, referred to as SD‐tree, for storing and sampling incident radiance, and presents a principled way to automatically budget training and rendering computations to minimize the variance of the final image.
Pre‐convolved Radiance Caching
The approximate interactive global illumination approach is an order of magnitude faster than Radiance Caching with Phong BRDFs and can be combined with Monte Carlo‐raytracing, Point‐based Global Illumination or Instant Radiosity.
Real-time global illumination by precomputed local reconstruction from sparse radiance probes
We present a direct-to-indirect transport technique that enables accurate real-time rendering of indirect illumination in mostly static scenes of complexity on par with modern games while supporting
Second-Order Occlusion-Aware Volumetric Radiance Caching
A second-order gradient analysis of light transport in participating media is presented and an improved radiance caching algorithm for volumetric light transport is developed, demonstrating that this method outperforms previous radiance cached algorithms both in terms of accurate derivative estimates and final radiance extrapolation.
View-dependent Radiance Caching
This work proposes a new radiance caching approach capable of more correctly reconstructing inter-reflections between glossy surfaces, all while improving performance compared to previous approaches.
Micro-rendering for scalable, parallel final gathering
This work presents an efficient and scalable method to compute global illumination solutions at interactive rates for complex and dynamic scenes, based on parallel final gathering running entirely on the GPU and demonstrates the applicability of the method to interactive global illumination, the simulation of multiple indirect bounces, and to final gathering from photon maps.
Clustered Pre-convolved Radiance Caching
A scalable method for rendering indirect illumination in diffuse and glossy scenes based on pre-convolved radiance caching (RC), which enables reusing the incident radiance computed at a surface point for its neighborhood using a pre-filtered voxel representation.
Global illumination with radiance regression functions
The approach is to exploit the nonlinear coherence of the indirect illumination data to make the RRF both compact and fast to evaluate, which enables real-time rendering with full global illumination effects, including changing caustics and multiple-bounce high-frequency glossy interreflections.
Real-time Radiance Caching using Chrominance Compression
This paper introduces the idea of expressing the radiance field in luminance/chrominance values and encoding the directional chrominance in lower detail. Reducing the spherical harmonics coefficients