-
Radiative Backpropagation with Non-Static Geometry
Authors:
Markus Worchel,
Ugo Finnendahl,
Marc Alexa
Abstract:
Radiative backpropagation-based (RB) methods efficiently compute reverse-mode derivatives in physically-based differentiable rendering by simulating the propagation of differential radiance. A key assumption is that differential radiance is transported like normal radiance. We observe that this holds only when scene geometry is static and demonstrate that current implementations of radiative backp…
▽ More
Radiative backpropagation-based (RB) methods efficiently compute reverse-mode derivatives in physically-based differentiable rendering by simulating the propagation of differential radiance. A key assumption is that differential radiance is transported like normal radiance. We observe that this holds only when scene geometry is static and demonstrate that current implementations of radiative backpropagation produce biased gradients when scene parameters change geometry. In this work, we derive the differential transport equation without assuming static geometry. An immediate consequence is that the parameterization matters when the sampling process is not differentiated: only surface integrals allow a local formulation of the derivatives, i.e., one in which moving surfaces do not affect the entire path geometry. While considerable effort has been devoted to handling discontinuities resulting from moving geometry, we show that a biased interior derivative compromises even the simplest inverse rendering tasks, regardless of discontinuities. An implementation based on our derivation leads to systematic convergence to the reference solution in the same setting and provides unbiased RB interior derivatives for path-space differentiable rendering.
△ Less
Submitted 2 July, 2025; v1 submitted 8 April, 2025;
originally announced April 2025.
-
Differentiable Shadow Mapping for Efficient Inverse Graphics
Authors:
Markus Worchel,
Marc Alexa
Abstract:
We show how shadows can be efficiently generated in differentiable rendering of triangle meshes. Our central observation is that pre-filtered shadow mapping, a technique for approximating shadows based on rendering from the perspective of a light, can be combined with existing differentiable rasterizers to yield differentiable visibility information. We demonstrate at several inverse graphics prob…
▽ More
We show how shadows can be efficiently generated in differentiable rendering of triangle meshes. Our central observation is that pre-filtered shadow mapping, a technique for approximating shadows based on rendering from the perspective of a light, can be combined with existing differentiable rasterizers to yield differentiable visibility information. We demonstrate at several inverse graphics problems that differentiable shadow maps are orders of magnitude faster than differentiable light transport simulation with similar accuracy -- while differentiable rasterization without shadows often fails to converge.
△ Less
Submitted 21 August, 2023;
originally announced August 2023.
-
Multi-View Mesh Reconstruction with Neural Deferred Shading
Authors:
Markus Worchel,
Rodrigo Diaz,
Weiwen Hu,
Oliver Schreer,
Ingo Feldmann,
Peter Eisert
Abstract:
We propose an analysis-by-synthesis method for fast multi-view 3D reconstruction of opaque objects with arbitrary materials and illumination. State-of-the-art methods use both neural surface representations and neural rendering. While flexible, neural surface representations are a significant bottleneck in optimization runtime. Instead, we represent surfaces as triangle meshes and build a differen…
▽ More
We propose an analysis-by-synthesis method for fast multi-view 3D reconstruction of opaque objects with arbitrary materials and illumination. State-of-the-art methods use both neural surface representations and neural rendering. While flexible, neural surface representations are a significant bottleneck in optimization runtime. Instead, we represent surfaces as triangle meshes and build a differentiable rendering pipeline around triangle rasterization and neural shading. The renderer is used in a gradient descent optimization where both a triangle mesh and a neural shader are jointly optimized to reproduce the multi-view images. We evaluate our method on a public 3D reconstruction dataset and show that it can match the reconstruction accuracy of traditional baselines and neural approaches while surpassing them in optimization runtime. Additionally, we investigate the shader and find that it learns an interpretable representation of appearance, enabling applications such as 3D material editing.
△ Less
Submitted 8 December, 2022;
originally announced December 2022.
-
Accurate Human Body Reconstruction for Volumetric Video
Authors:
Decai Chen,
Markus Worchel,
Ingo Feldmann,
Oliver Schreer,
Peter Eisert
Abstract:
In this work, we enhance a professional end-to-end volumetric video production pipeline to achieve high-fidelity human body reconstruction using only passive cameras. While current volumetric video approaches estimate depth maps using traditional stereo matching techniques, we introduce and optimize deep learning-based multi-view stereo networks for depth map estimation in the context of professio…
▽ More
In this work, we enhance a professional end-to-end volumetric video production pipeline to achieve high-fidelity human body reconstruction using only passive cameras. While current volumetric video approaches estimate depth maps using traditional stereo matching techniques, we introduce and optimize deep learning-based multi-view stereo networks for depth map estimation in the context of professional volumetric video reconstruction. Furthermore, we propose a novel depth map post-processing approach including filtering and fusion, by taking into account photometric confidence, cross-view geometric consistency, foreground masks as well as camera viewing frustums. We show that our method can generate high levels of geometric detail for reconstructed human bodies.
△ Less
Submitted 26 February, 2022;
originally announced February 2022.
