Antithetic sampling for Monte Carlo differentiable rendering

Antithetic sampling for Monte Carlo differentiable rendering
复制标题

蒙特卡罗可微分渲染的对偶采样

DOI:
10.1145/3450626.3459783
复制
发表时间:
2021
影响因子:
6.2
通讯作者:
Zhao, Shuang
Zhao, Shuang
中科院分区:
计算机科学1区
文献类型:
--
作者:
Zhang, Cheng;Dong, Zhao;Doggett, Michael;Zhao, Shuang

文献摘要

参考文献

被引文献

相似文献

光传输路径的随机采样是蒙特卡罗前向渲染的关键,之前的研究已经形成了能够在复杂场景中绘制高贡献光路的成熟技术。这些采样技术也已应用于可微分渲染。在本文中,我们证明了为前向渲染开发的路径采样技术对于光泽材质的可微分渲染可能会变得低效,尤其是在估计相对于全局场景几何形状的导数时。为了解决这个问题,我们引入了 BSDF 和光传输路径的对立采样,从而可以显着加快收敛速度​​,并且可以轻松集成到现有的可微渲染管道中。我们通过将我们的导数估计与现有无偏技术生成的估计进行比较来验证我们的方法。此外,我们通过与现有采样方法进行同等质量和同等时间的比较来证明我们技术的有效性。
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 introduceantithetic samplingof 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.
梯度域路径追踪
DOI: 10.1145/2766997
发表时间: 2015
期刊: ACM Transactions on Graphics (TOG)
影响因子: --
作者:
M. Kettunen;Marco Manzi;M. Aittala;J. Lehtinen;F. Durand;Matthias Zwicker
通讯作者: Matthias Zwicker
DOI: 10.1017/s0305004100031467
发表时间: 1956
影响因子: 0.8
作者:
J. Hammersley;J. Mauldon
通讯作者: J. Mauldon
DOI: 10.1111/cgf.12416
发表时间: 2014
影响因子: 2.5
作者:
Kartic Subr;D. Nowrouzezahrai;Wojciech Jarosz;J. Kautz;Kenny Mitchell
通讯作者: Kenny Mitchell