Research map / 3D vision / NeRF and radiance fields
Complementary Intrinsics From Neural Radiance Fields and CNNs for Outdoor Scene Relighting
Siqi Yang, Xuanning Cui, Yongjie Zhu, Jiajun Tang, Si Li, Zhaofei Yu, Boxin Shi
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | PSNR: 23.4755 -> 24.2387 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | MSE: 0.0051 -> 0.0038 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | SSIM: 0.8595 -> 0.8558 | Hurt |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | LPIPS: 0.1336 -> 0.1354 | Hurt |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | MAE: 0.0543 -> 0.0430 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | PSNR: 17.3381 -> 24.2387 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | SSIM: 0.5999 -> 0.8558 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | MSE: 0.0204 -> 0.0038 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | MAE: 0.1176 -> 0.0430 | Improved |
| Use NeRF-produced intrinsic pseudo-labels to train the single-image CNN | Outdoor scene intrinsic decomposition · Outdoor landmark image collections, test set | LPIPS: 0.1512 -> 0.1354 | Improved |
| Separate sky rendering from opaque-object photometric rendering using SkyMLP (ablation) | Outdoor scene relighting · Outdoor landmark scenes | Visual sky realism | Improved |
| Use the hybrid object-sky image formation and combine it with the uninterpretable NeRF-W rendering (ablation) | NeRF intrinsic estimation · Outdoor landmark scenes | Shadow and geometry quality | Improved |
Closest papers by meaning
- NeRF for Outdoor Scene Relighting (ECCV 2022 · similarity 0.92)
- ReNeRF: Relightable Neural Radiance Fields with Nearfield Lighting (ICCV 2023 · similarity 0.90)
- Self-supervised Outdoor Scene Relighting (ECCV 2020 · similarity 0.90)
- Spatially-Varying Outdoor Lighting Estimation From Intrinsics (CVPR 2021 · similarity 0.89)
- Neural Fields Meet Explicit Geometric Representations for Inverse Rendering of Urban Scenes (CVPR 2023 · similarity 0.89)
- A Pre-convolved Representation for Plug-and-Play Neural Illumination Fields (AAAI 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
