tkuri/Awesome-InverseRendering
tkuri/Awesome-InverseRendering has 236 GitHub stars and 11 forks. Star changes are not measured yet. Last checked 2026-11-15 19:26 UTC. Last push 2026-10-10.
Measured star growth
- in 24h
- Collecting history
- in 7 days
- Collecting history
- in 15 days
- Collecting history
- in 1 month
- Collecting history
Net star changes over overlapping periods, not daily totals. Missing measurements are left blank. Stars measure attention, not software quality.
About
List of awesome papers on Intrinsic Decomposition & Inverse Rendering
Creative & MediaREADME.md
OriginalIntrinsic Decomposition, Inverse Rendering & Portrait Relighting

A curated reading list of intrinsic image decomposition, recovery of scene geometry/materials/lighting, and portrait relighting.
Scope: papers published at major vision, graphics and machine-learning conferences or journals, including CVPR, ICCV, ECCV, SIGGRAPH / SIGGRAPH Asia, NeurIPS, ICLR, AAAI, TPAMI, IJCV, TOG and CGF. 3DV, WACV, Eurographics / EGSR, and relevant workshops of these conferences are also included. Workshop papers are explicitly labeled. An arXiv link may point to the author version of an accepted paper; arXiv-only work is excluded.
DL means that the method uses deep learning: a trained network, a scene-optimized neural field/MLP, or a pretrained neural prior. non-DL covers explicit physical models and parameter optimization without such a component; automatic differentiation or Gaussian Splatting alone does not imply DL. Hybrid methods are placed in DL when their reconstruction or relighting pipeline uses neural components. Auxiliary masking/feature matching alone is not used to assign DL.
Years refer to the listed publication venue, rather than the first preprint. Resources are official code, data or author project pages where identified; an empty cell means no resource is linked here. Related papers are listed once under their main task. Dataset papers may also occur in method tables when the dataset accompanies a method.
Last literature audit: 2026-10-11. See research notes for search coverage, corrections and remaining limitations.
Table of contents
- Surveys
- Intrinsic Decomposition (non-DL)
- Intrinsic Decomposition (DL)
- Inverse Rendering (non-DL)
- Inverse Rendering (DL)
- Portrait Relighting (DL)
- Dataset
Surveys
| Year | Pub | Paper | Resources |
|---|---|---|---|
| 2017 | Eurographics / CGF | Intrinsic Decompositions for Image Editing | Code Dataset |
| 2021 | CGF | Deep Neural Models for Illumination Estimation and Relighting: A Survey | |
| 2022 | IJCV | A Survey on Intrinsic Images: Delving Deep Into Lambert and Beyond | |
| 2026 | CGF | Recent Trends in Inverse Rendering |
Intrinsic Decomposition (non-DL)
| Year | Pub | Paper | Resources |
|---|---|---|---|
| 2001 | ICCV | Deriving Intrinsic Images from Image Sequences | |
| 2005 | TPAMI | Recovering Intrinsic Images from a Single Image | |
| 2008 | CVPR | Intrinsic Image Decomposition with Non-Local Texture Cues | |
| 2009 | SIGGRAPH Asia | User-Assisted Intrinsic Images | |
| 2011 | NeurIPS | Recovering Intrinsic Images with a Global Sparsity Prior on Reflectance | |
| 2012 | EGSR / CGF | Intrinsic Images by Clustering | |
| 2016 | SIGGRAPH | Live Intrinsic Video | |
| 2021 | Eurographics | Interactive Photo Editing on Smartphones via Intrinsic Decomposition | |
| 2021 | SIGGRAPH | Real-time Global Illumination Decomposition of Videos | |
| 2023 | CVPR | Light Source Separation and Intrinsic Image Decomposition Under AC Illumination |
Intrinsic Decomposition (DL)
Inverse Rendering (non-DL)
L: Lighting Estimation, G: Geometry Estimation, M: Material Estimation. In the legacy tables, a check indicates an estimated quantity; an empty cell does not assert that the quantity is absent from the pipeline. New entries describe the capture setting directly.
Material capture
| Year | Pub | Paper | Resources | L | G | M |
|---|---|---|---|---|---|---|
| 2010 | SIGGRAPH | Manifold bootstrapping for SVBRDF capture | ✓ | ✓ | ||
| 2013 | SIGGRAPH | Practical SVBRDF Capture in the Frequency Domain | ✓ | ✓ | ||
| 2014 | SIGGRAPH ASIA | Appearance-from-Motion: Recovering Spatially Varying Surface Reflectance under Unknown Lighting | ✓ | ✓ | ✓ | |
| 2015 | CGF | Mobile Surface Reflectometry | ✓ | ✓ | ||
| 2015 | SIGGRAPH | Two-Shot SVBRDF Capture for Stationary Materials | Code & Dataset | ✓ | ✓ | |
| 2016 | SIGGRAPH ASIA | Recovering Shape and Spatially-Varying Surface Reflectance under Unknown Illumination | ✓ | ✓ | ✓ | |
| 2016 | SIGGRAPH ASIA | Simultaneous Acquisition of Microscale Reflectance and Normals | ✓ | ✓ | ||
| 2017 | SIGGRAPH ASIA | Polarization imaging reflectometry in the wild | ✓ | ✓ | ||
| 2018 | SIGGRAPH ASIA | Simultaneous Acquisition of Polarimetric SVBRDF and Normals | ✓ | ✓ | ||
| 2021 | EGSR | Material and Lighting Reconstruction for Complex Indoor Scenes with Texture-space Differentiable Rendering | ✓ | ✓ | ||
| 2021 | Eurographics | Progressive Acquisition of SVBRDF and Shape in Motion | ✓ | ✓ |
Physical models and differentiable optimization
| Year | Pub | Paper | Resources | Focus / capture setting |
|---|---|---|---|---|
| 1999 | SIGGRAPH | Inverse Global Illumination: Recovering Reflectance Models of Real Scenes from Photographs | Scene reflectance and global illumination | |
| 2015 | TPAMI | Shape, Illumination, and Reflectance from Shading | SIRFS; single-image shape, reflectance and lighting | |
| 2020 | ECCV | Polarimetric Multi-View Inverse Rendering | Multi-view geometry/material optimization with polarization | |
| 2021 | EGSR / CGF | Unified Shape and SVBRDF Recovery using Differentiable Monte Carlo Rendering | Joint mesh and SVBRDF optimization |
Explicit Gaussian inverse rendering
Inverse Rendering (DL)
Image-based estimation and material capture
Neural fields, multi-view reconstruction and reflectance
Image and video decomposition, diffusion and material priors
Lighting estimation and HDR environments
| Year | Pub | Paper | Resources | Focus / capture setting |
|---|---|---|---|---|
| 2020 | ECCV | PointAR: Efficient Lighting Estimation for Mobile Augmented Reality | RGB-D-conditioned mobile lighting estimation | |
| 2022 | ECCV | Editable Indoor Lighting Estimation | Parametric lights and high-frequency illumination texture | |
| 2022 | ECCV | StyleLight: HDR Panorama Generation for Lighting Estimation and Editing | Project | Single-image HDR panorama lighting estimation/editing |
| 2023 | ICCV | EverLight: Indoor-Outdoor Editable HDR Lighting Estimation | Editable indoor/outdoor HDR illumination | |
| 2024 | CVPR | LightOctree: Lightweight 3D Spatially-Coherent Indoor Lighting Estimation | Octree-based spatially varying indoor lighting | |
| 2026 | CVPR | Lighting in Motion: Spatiotemporal HDR Lighting Estimation | Spatiotemporal HDR environment lighting from diffusion |
Human and face inverse rendering
Gaussian methods with neural components
Scene and video relighting
| Year | Pub | Paper | Resources | Focus / capture setting |
|---|---|---|---|---|
| 2020 | ECCV | Learning to Factorize and Relight a City | Outdoor time-varying illumination and permanent factors | |
| 2020 | ECCV | Self-supervised Outdoor Scene Relighting | Self-supervised albedo, geometry, lighting and relighting | |
| 2021 | CVPR Workshops (NTIRE) | Physically Inspired Dense Fusion Networks for Relighting | Depth-guided image relighting with intrinsic decomposition | |
| 2022 | ECCV | NeRF for Outdoor Scene Relighting | NeRF-OSR; outdoor illumination and viewpoint editing | |
| 2023 | CVPR | Complementary Intrinsics From Neural Radiance Fields and CNNs for Outdoor Scene Relighting | Outdoor relighting; CNN and NeRF intrinsic estimates | |
| 2023 | CVPR | Weakly-Supervised Single-View Image Relighting | Project | Single-view image relighting with weak supervision |
| 2024 | WACV | Fast Sun-Aligned Outdoor Scene Relighting Based on TensoRF | Outdoor TensoRF relighting with sun alignment | |
| 2024 | ECCV | The Sky's the Limit: Relightable Outdoor Scenes via a Sky-pixel Constrained Illumination Prior and Outside-In Visibility | Code | Outdoor relighting; sky-constrained illumination and visibility |
| 2025 | ICCV | Light-A-Video: Training-free Video Relighting via Progressive Light Fusion | Pretrained diffusion models; temporally consistent relighting | |
| 2025 | ICCV | LightSwitch: Multi-view Relighting with Material-guided Diffusion | Multi-view material-guided diffusion relighting | |
| 2025 | CVPR | LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting | Latent intrinsics for indoor image relighting | |
| 2025 | SIGGRAPH | Physically Controllable Relighting of Photographs | Single-image relighting using intrinsic and geometric predictions | |
| 2026 | CVPR | FlowPortal: Residual-Corrected Flow for Training-Free Video Relighting and Background Replacement | Pretrained flow-model video relighting and background replacement |