论文标题
高原可区分的路径追踪
Plateau-reduced Differentiable Path Tracing
论文作者
论文摘要
当前可区分的渲染器提供有关任意场景参数的光传输梯度。但是，仅这些梯度的存在并不能保证优化中有用的更新步骤。取而代之的是，由于固有的高原（即零梯度的区域）在目标函数中，反向渲染可能不会收敛。我们建议通过浏览高维渲染函数来减轻这种情况，该函数将场景参数映射到模糊参数空间的附加内核中的图像。我们描述了两个蒙特卡洛估计器，以有效地计算无高原梯度，即具有较低的方差，并表明这些梯度在优化误差和运行时性能中转化为净收益。我们的方法是对黑框和可区分渲染器的直接扩展，并可以优化复杂的轻型传输（例如苛性或全球照明）的问题，现有可区分的渲染器不会融合。
Current differentiable renderers provide light transport gradients with respect to arbitrary scene parameters. However, the mere existence of these gradients does not guarantee useful update steps in an optimization. Instead, inverse rendering might not converge due to inherent plateaus, i.e., regions of zero gradient, in the objective function. We propose to alleviate this by convolving the high-dimensional rendering function that maps scene parameters to images with an additional kernel that blurs the parameter space. We describe two Monte Carlo estimators to compute plateau-free gradients efficiently, i.e., with low variance, and show that these translate into net-gains in optimization error and runtime performance. Our approach is a straightforward extension to both black-box and differentiable renderers and enables optimization of problems with intricate light transport, such as caustics or global illumination, that existing differentiable renderers do not converge on.
