Research map / 3D vision / NeRF and radiance fields
ActiveNeRF: Learning Where to See with Uncertainty Estimation
Xuran Pan, Zihang Lai, Shiji Song, Gao Huang
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 uncertainty estimation | NeRF · NeRF synthetic (10 images) | PSNR: 28.04 -> 28.51 | Improved |
| add uncertainty estimation | NeRF · NeRF synthetic (5 images) | PSNR: 21.14 -> 23.23 | Improved |
| add uncertainty estimation | NeRF · NeRF synthetic (all images) | PSNR: 31.01 -> 30.45 | No clear effect |
| add uncertainty estimation | NeRF · LLFF realistic (5 images) | PSNR: 21.67 -> 22.03 | Improved |
| add uncertainty estimation | NeRF · LLFF realistic (all images) | PSNR: 26.50 -> 25.96 | No clear effect |
| use ActiveNeRF acquisition instead of random view capture (continuous learning) | NeRF · NeRF synthetic (20 observations) | PSNR: 24.25 -> 26.24 | Improved |
| use ActiveNeRF acquisition instead of furthest view sampling (continuous learning) | NeRF · NeRF synthetic (20 observations) | PSNR: 26.00 -> 26.24 | Improved |
| use ActiveNeRF acquisition instead of random view capture (continuous learning) | NeRF · LLFF realistic (10 observations) | PSNR: 18.49 -> 20.14 | Improved |
| use continuous learning instead of Bayesian estimation | ActiveNeRF · NeRF synthetic (20 observations) | PSNR: 25.67 -> 26.24 | Improved |
| use Bayesian estimation instead of continuous learning | ActiveNeRF · NeRF synthetic (20 observations) | time: 2.2h -> 30min | Improved |
| add acquisition function over random selection (ablation) | ActiveNeRF · NeRF synthetic (20 observations) | PSNR: 24.77 -> 26.24 | Improved |
| add acquisition function over furthest view sampling (ablation) | ActiveNeRF · NeRF synthetic (20 observations) | PSNR: 26.05 -> 26.24 | Improved |
Later papers that compared with this method
4 later accepted papers report ActiveNeRF as a baseline in their results tables; 4 of them report a better number on most of the results they share with it.
- FluidGaussian: Propagating Simulation-Based Uncertainty Toward Functionally-Intelligent 3D Reconstruction (CVPR 2026 · better on 8 of 8 shared results)
- GO-PRE: Goal-Oriented Next-Best-View Selection via Predictive Rendering Entropy for Active 3D Reconstruction (ICML 2026 · better on 1 of 1 shared result)
- GURecon: Learning Detailed 3D Geometric Uncertainties for Neural Surface Reconstruction (AAAI 2025 · better on 6 of 6 shared results)
- FisherRF: Active View Selection and Mapping with Radiance Fields using Fisher Information (ECCV 2024 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Neural Visibility Field for Uncertainty-Driven Active Mapping (CVPR 2024 · similarity 0.91)
- AligNeRF: High-Fidelity Neural Radiance Fields via Alignment-Aware Training (CVPR 2023 · similarity 0.91)
- ProvNeRF: Modeling per Point Provenance in NeRFs as a Stochastic Field (NeurIPS 2024 · similarity 0.91)
- FisherRF: Active View Selection and Mapping with Radiance Fields using Fisher Information (ECCV 2024 · similarity 0.90)
- SinNeRF: Training Neural Radiance Fields on Complex Scenes from a Single Image (ECCV 2022 · similarity 0.90)
- ColNeRF: Collaboration for Generalizable Sparse Input Neural Radiance Field (AAAI 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.
