Research map / 3D vision / NeRF and radiance fields
Grid-Guided Neural Radiance Fields for Large Urban Scenes
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Combine ground feature planes with a jointly trained NeRF branch | Grid-guided NeRF · Rubble, Campus, Residential | PSNR | Helped |
| Add grid features as inputs to NeRF without tuning the grid (ablation) | NeRF branch · Large urban scenes | PSNR: ∼1db | Helped |
| Tune the grid features jointly with NeRF (ablation) | NeRF branch with grid features · Large urban scenes | PSNR: ∼2.5db | Helped |
| Add NeRF supervision to a high-resolution grid (ablation) | Grid branch · Campus | PSNR: ∼2db | Helped |
| Use a single low-resolution feature grid (ablation) | Grid branch · Large urban scenes | rendering quality | Hurt |
| Use a single high-resolution feature grid (ablation) | Grid branch · Large urban scenes | rendering quality | Hurt |
| Reduce NeRF capacity to D=3, W=32 (ablation) | NeRF branch · Large urban scenes | rendering quality | Hurt |
| Increase PE frequency while keeping NeRF capacity small (ablation) | Small NeRF branch · Large urban scenes | rendering quality | No clear effect |
| Increase NeRF capacity to D=3, W=256 (ablation) | NeRF branch · Large urban scenes | rendering quality | Helped |
| Increase PE frequency with sufficient NeRF capacity (ablation) | NeRF branch · Large urban scenes | rendering quality | Helped |
| Use the NeRF branch to refine grid features (ablation) | Grid branch · Large urban scenes | rendering quality | Helped |
| Use a compact ground-plane and shared vertical-vector representation | Grid-guided NeRF · Large urban scenes | PSNR: 21.075 vs. 20.915 | No clear effect |
Closest papers by meaning
- NeRFLight: Fast and Light Neural Radiance Fields Using a Shared Feature Grid (CVPR 2023 · similarity 0.93)
- Learning Robust Generalizable Radiance Field with Visibility and Feature Augmented Point Representation (ICLR 2024 · similarity 0.90)
- Multi-Level Neural Scene Graphs for Dynamic Urban Environments (CVPR 2024 · similarity 0.90)
- Dynamic 3D Gaussian Fields for Urban Areas (NeurIPS 2024 · similarity 0.90)
- Mask-Based Modeling for Neural Radiance Fields (ICLR 2024 · similarity 0.90)
- Is Vanilla MLP in Neural Radiance Field Enough for Few-shot View Synthesis? (CVPR 2024 · similarity 0.89)
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.