Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields
Ref-NeRF: Structured View-Dependent Appearance for Neural Radiance Fields
复制标题
DOI：
10.1109/cvpr52688.2022.00541
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Dor Verbin;Peter Hedman;B. Mildenhall;Todd E. Zickler;J. Barron;Pratul P. Srinivasan
中科院分区：
文献类型：
--
作者：
Dor Verbin;Peter Hedman;B. Mildenhall;Todd E. Zickler;J. Barron;Pratul P. Srinivasan
神经辐射场(NERF)是一种流行的视图合成技术，它将场景表示为连续的体积函数，并由多层感知器进行参数化，这些感知器提供每个位置的体积密度和与视图相关的发射辐射。虽然基于NERF的技术擅长表示具有平滑变化的视相关外观的精细几何结构，但它们往往无法准确地捕捉和重现光滑表面的外观。我们通过引入Ref-Nerf来解决这一限制，它将Nerf对依赖于视图的出射辐射的参数化替换为反射辐射的表示，并使用一组空间变化的场景属性来构造该函数。我们表明，结合法线向量的正则化，我们的模型显著提高了镜面反射的真实感和准确性。此外，我们还表明，我们的模型的出射辐射度的内部表示是可解释的，并且对于场景编辑是有用的。
Neural Radiance Fields (NeRF) is a popular view synthesis technique that represents a scene as a continuous volumetric function, parameterized by multilayer perceptrons that provide the volume density and view-dependent emitted radiance at each location. While NeRF-based techniques excel at representing fine geometric structures with smoothly varying view-dependent appearance, they often fail to accurately capture and reproduce the appearance of glossy surfaces. We address this limitation by introducing Ref-NeRF, which replaces NeRF's parameterization of view-dependent outgoing radiance with a representation of reflected radiance and structures this function using a collection of spatially-varying scene properties. We show that together with a regularizer on normal vectors, our model significantly improves the realism and accuracy of specular reflections. Furthermore, we show that our model's internal representation of outgoing radiance is interpretable and useful for scene editing.