PureLight: Learning Complex Luminaires with Light Tracing

2026-06-02

AI Summary

This work addresses the challenge of efficiently rendering complex light sources—such as emitters enclosed by multiple reflective layers—whose intricate light transport defies conventional path tracing. The authors propose a light-tracing-based appearance modeling approach that represents outgoing radiance as a probability distribution and, for the first time, employs a normalizing flow network to learn its density function. Physical correctness is ensured by incorporating flux constraints to recover absolute radiance values. To enable efficient inference, knowledge distillation compresses the normalizing flow into a lightweight MLP. Additionally, a dedicated sampling and mixing network is introduced to support direct illumination computation and scene composition. The method achieves high-quality rendering of traditionally intractable complex light sources across arbitrary scenes using only a small number of samples.

Abstract

We propose a neural formulation for estimating the appearance of complex luminaires. We focus on challenging luminaires with complex light transport (e.g., small emitters enclosed by multiple specular layers) that are difficult for (bidirectional) path tracing. To this end, we use light tracing to construct paths from emitters to the exit surfaces and formulate appearance estimation as a distribution learning problem. Specifically, we model the probability density function (pdf) of outgoing radiance on the exit surfaces using a large normalizing flow network, and recover the outgoing radiance as the product of the estimated pdf and flux. To enable efficient inference, we distill the learned appearance into a lightweight MLP that directly estimates radiance on the exit surfaces. We additionally train a sampling network for effective direct illumination computation from the luminaire, and a blending network to composite the luminaire into the scene. Our formulation makes it feasible to render challenging luminaires using low sample counts in arbitrary scenes.

Problem

• complex luminaires
• light transport
• appearance estimation
• rendering
• specular layers

Innovation

• light tracing
• normalizing flow
• neural rendering
• distribution learning
• luminaire appearance