# Inverse Global Illumination using a Neural Radiometric Prior

arXiv: https://arxiv.org/abs/2305.02192 · HTML: https://mlsanity.com/paper/2305.02192/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 16 of its 43 references that are in the map.
- First posted on arXiv 3 May 2023.

## Builds on (most cited first, 16 of 16)

- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [NeRV: Neural Reflectance and Visibility Fields for Relighting and View Synthesis](https://mlsanity.com/paper/2012.03927/) (2020)
- [Modular Primitives for High-Performance Differentiable Rendering](https://mlsanity.com/paper/2011.03277/) (2020)
- [Neural 3D Mesh Renderer](https://mlsanity.com/paper/1711.07566/) (2017)
- [Advances in Neural Rendering](https://mlsanity.com/paper/2111.05849/) (2021)
- [State of the Art on Neural Rendering](https://mlsanity.com/paper/2004.03805/) (2020)
- [Modeling Indirect Illumination for Inverse Rendering](https://mlsanity.com/paper/2204.06837/) (2022)
- [NeRFactor: Neural Factorization of Shape and Reflectance Under an Unknown Illumination](https://mlsanity.com/paper/2106.01970/) (2021)
- [Real-time Neural Radiance Caching for Path Tracing](https://mlsanity.com/paper/2106.12372/) (2021)
- [A Versatile Scene Model with Differentiable Visibility Applied to Generative Pose Estimation](https://mlsanity.com/paper/1602.03725/) (2016)
- [Neural Control Variates](https://mlsanity.com/paper/2006.01524/) (2020)
- [Neural Radiosity](https://mlsanity.com/paper/2105.12319/) (2021)
- [Soft Rasterizer: Differentiable Rendering for Unsupervised Single-View Mesh Reconstruction](https://mlsanity.com/paper/1901.05567/) (2019)
- [Pix2Vex: Image-to-Geometry Reconstruction using a Smooth Differentiable Renderer](https://mlsanity.com/paper/1903.11149/) (2019)
- [Differentiable Neural Radiosity](https://mlsanity.com/paper/2201.13190/) (2022)

## Built on by (most cited first, 9 of 9)

- [NeISF: Neural Incident Stokes Field for Geometry and Material Estimation](https://mlsanity.com/paper/2311.13187/) (2023)
- [Differentiable Light Transport with Gaussian Surfels via Adapted Radiosity for Efficient Relighting and Geometry Reconstruction](https://mlsanity.com/paper/2509.18497/) (2025)
- [GaNI: Global and Near Field Illumination Aware Neural Inverse Rendering](https://mlsanity.com/paper/2403.15651/) (2024)
- [ESR-NeRF: Emissive Source Reconstruction Using LDR Multi-view Images](https://mlsanity.com/paper/2404.15707/) (2024)
- [Radiometrically Consistent Gaussian Surfels for Inverse Rendering](https://mlsanity.com/paper/2603.01491/) (2026)
- [Neural Inverse Rendering from Propagating Light](https://mlsanity.com/paper/2506.05347/) (2025)
- [Fast and Accurate Neural Rendering Using Semi-Gradients](https://mlsanity.com/paper/2410.10149/) (2024)
- [GLOW: Global Illumination-Aware Inverse Rendering of Indoor Scenes Captured with Dynamic Co-Located Light & Camera](https://mlsanity.com/paper/2511.22857/) (2025)
- [Rotated Lights for Consistent and Efficient 2D Gaussians Inverse Rendering](https://mlsanity.com/paper/2602.08724/) (2026)

## Abstract (arXiv)

Inverse rendering methods that account for global illumination are becoming more popular, but current methods require evaluating and automatically differentiating millions of path integrals by tracing multiple light bounces, which remains expensive and prone to noise. Instead, this paper proposes a radiometric prior as a simple alternative to building complete path integrals in a traditional differentiable path tracer, while still correctly accounting for global illumination. Inspired by the Neural Radiosity technique, we use a neural network as a radiance function, and we introduce a prior consisting of the norm of the residual of the rendering equation in the inverse rendering loss. We train our radiance network and optimize scene parameters simultaneously using a loss consisting of both a photometric term between renderings and the multi-view input images, and our radiometric prior (the residual term). This residual term enforces a physical constraint on the optimization that ensures that the radiance field accounts for global illumination. We compare our method to a vanilla differentiable path tracer, and more advanced techniques such as Path Replay Backpropagation. Despite the simplicity of our approach, we can recover scene parameters with comparable and in some cases better quality, at considerably lower computation times.