Path Guiding Using Spatio-Directional Mixture Models
This work addresses path guiding in rendering for computer graphics, offering a novel approach to improve accuracy and efficiency in scenes with complex lighting, though it appears incremental as it builds on existing mixture model methods.
The paper tackles the challenge of constructing light paths in path tracing by proposing a learning-based method that uses spatio-directional Gaussian mixture models (SDMMs) to approximate incident radiance and BSDFs, addressing issues like capturing spatial-directional correlations and enabling approximate product sampling with arbitrary BSDFs. It performs well on scenes with small, localized luminaires, showing improved efficiency in handling high spatio-directional correlation.
We propose a learning-based method for light-path construction in path tracing algorithms, which iteratively optimizes and samples from what we refer to as spatio-directional Gaussian mixture models (SDMMs). In particular, we approximate incident radiance as an online-trained $5$D mixture that is accelerated by a $k$D-tree. Using the same framework, we approximate BSDFs as pre-trained $n$D mixtures, where $n$ is the number of BSDF parameters. Such an approach addresses two major challenges in path-guiding models. First, the $5$D radiance representation naturally captures correlation between the spatial and directional dimensions. Such correlations are present in e.g. parallax and caustics. Second, by using a tangent-space parameterization of Gaussians, our spatio-directional mixtures can perform approximate product sampling with arbitrarily oriented BSDFs. Existing models are only able to do this by either foregoing anisotropy of the mixture components or by representing the radiance field in local (normal aligned) coordinates, which both make the radiance field more difficult to learn. An additional benefit of the tangent-space parameterization is that each individual Gaussian is mapped to the solid sphere with low distortion near its center of mass. Our method performs especially well on scenes with small, localized luminaires that induce high spatio-directional correlation in the incident radiance.