Research map / Point clouds / Point cloud classification
VF-NeRF: Viewshed Fields for Rigid NeRF Registration
Leo Segre, Shai Avidan
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 |
|---|---|---|---|
| add viewshed-field-based photometric initialization to iNeRF | iNeRF · LLFF full overlap | translation error Δt (x1e2): 22.8907 -> 0.2254 | Improved |
| replace iNeRF with VF-NeRF view and ray sampling | photometric registration · LLFF full overlap | translation error Δt (x1e2): 0.2254 -> 0.0151 | Improved |
| refine VF point cloud FPFH+RANSAC init with VF-NeRF optimization | VF-NeRF · LLFF full overlap | translation error Δt (x1e2): 2.1149 -> 0.0162 | Improved |
| add T0 initialization (ablation) | VF-NeRF · LLFF | translation error Δt (x1e2): 39.2993 -> 0.0151 | Improved |
| add viewshed masks for ray sampling (ablation) | VF-NeRF · LLFF | translation error Δt (x1e2): 1.7348 -> 0.0151 | Improved |
| optimize over a set of novel views instead of a single image (ablation) | VF-NeRF · LLFF | translation error Δt (x1e2): 9.4008 -> 0.0151 | Improved |
| add up to 20% noise to oriented point positions (ablation) | VF-NeRF · LLFF Trex | translation error Δt (x1e2): 0.0113 -> 0.0149 | No clear effect |
| use original images instead of VF-generated views (ablation) | photometric registration · LLFF full overlap | translation error Δt (x1e2): 0.0151 -> 0.0132 | Mixed |
| VF-NeRF with PC init vs best baseline on Objaverse (all scenes) | VF-NeRF · Objaverse (100% scenes) | rotation error ΔR: 9.65 -> 6.77 | Improved |
| symmetric scenes can converge to a solution 180 degrees off | VF-NeRF · Objaverse | rotation error | Hurt |
Closest papers by meaning
- AligNeRF: High-Fidelity Neural Radiance Fields via Alignment-Aware Training (CVPR 2023 · similarity 0.91)
- DReg-NeRF: Deep Registration for Neural Radiance Fields (ICCV 2023 · similarity 0.90)
- Joint Optimization of Neural Radiance Fields and Continuous Camera Motion from a Monocular Video (CVPR 2025 · similarity 0.88)
- CF-NeRF: Camera Parameter Free Neural Radiance Fields with Incremental Learning (AAAI 2024 · similarity 0.88)
- ProvNeRF: Modeling per Point Provenance in NeRFs as a Stochastic Field (NeurIPS 2024 · similarity 0.88)
- NerfBaselines: Consistent and Reproducible Evaluation of Novel View Synthesis Methods (NeurIPS 2025 · similarity 0.88)
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.
