Antithetic Sampling for Monte Carlo Differentiable Rendering
Cheng Zhang1,2, Zhao Dong2, Michael Doggett3,2, and Shuang Zhao1
1University of California, Irvine     2Facebook Reality Labs     3Lund University
ACM Transactions on Graphics (SIGGRAPH 2021), 40(4), 2021
teaser
Abstract

Stochastic sampling of light transport paths is key to Monte Carlo forward rendering, and previous studies have led to mature techniques capable of drawing high-contribution light paths in complex scenes. These sampling techniques have also been applied to differentiable rendering.

In this paper, we demonstrate that path sampling techniques developed for forward rendering can become inefficient for differentiable rendering of glossy materials---especially when estimating derivatives with respect to global scene geometries. To address this problem, we introduce antithetic sampling of BSDFs and light-transport paths, allowing significantly faster convergence and can be easily integrated into existing differentiable rendering pipelines. We validate our method by comparing our derivative estimates to those generated with existing unbiased techniques. Further, we demonstrate the effectiveness of our technique by providing equal-quality and equal-time comparisons with existing sampling methods.

Downloads
  • Paper: pdf (12 MB)
  • Supplemental material: html, zip (59 MB)
  • Code: coming soon!
Bibtex citation
@article{Zhang:2021:Antithetic,
  title={Antithetic Sampling for Monte Carlo Differentiable Rendering},
  author={Zhang, Cheng and Dong, Zhao and Doggett, Michael and Zhao, Shuang},
  journal={ACM Trans. Graph.},
  volume={40},
  number={4},
  year={2021},
  pages={77:1--77:12}
}
Acknowledgments

We thank the anonymous reviewers for their constructive comments. Cheng and Shuang are partially supported by NSF grant 1900927. Michael is partially supported by ELLIIT and the Wallenberg AI, Autonomous Systems and Software Program (WASP) funded by the Knut and Alice Wallenberg Foundation.