Research map / 3D vision / NeRF and radiance fields
FisherRF: Active View Selection and Mapping with Radiance Fields using Fisher Information
Wen Jiang, BOSHU LEI, Kostas Daniilidis
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 |
|---|---|---|---|
| select views by Fisher Information EIG instead of random on 3D Gaussian Splatting | 3D Gaussian Splatting · Blender (sequential, 20 views) | PSNR: 28.732 -> 29.525 | Improved |
| select views by EIG instead of ActiveNeRF variance selection | 3D Gaussian Splatting · Blender (sequential, 20 views) | PSNR: 26.610 -> 29.525 | Improved |
| batch EIG selection instead of random | 3D Gaussian Splatting · Blender (batch of 4) | PSNR: 27.135 -> 29.094 | Improved |
| EIG selection instead of random with only 10 views | 3D Gaussian Splatting · Blender (10 views) | PSNR: 22.493 -> 23.681 | Improved |
| EIG selection instead of random on real-world scenes | 3D Gaussian Splatting · Mip-NeRF360 | PSNR: 17.914 -> 20.351 | Improved |
| use raw uncertainty (BayesRays) for view selection instead of random | Nerfacto · Blender (sequential, 20 views) | PSNR: 22.614 -> 22.504 | Hurt |
| use ActiveNeRF variance selection instead of random | Plenoxels · Blender (sequential, 20 views) | PSNR: 23.906 -> 23.522 | Hurt |
| use ActiveNeRF selection instead of random | 3D Gaussian Splatting · Mip-NeRF360 (sequential) | PSNR: 17.914 -> 17.889 | No clear effect |
| EIG-based active mapping vs prior active mapping methods | SplaTAM-based active mapping · Gibson | Completion (%): 80.45 -> 92.89 | Improved |
| Fisher-based pixel uncertainty vs BayesRays | 3D Gaussian Splatting · LF dataset | AUSE (average): 0.23 -> 0.22 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- CF-NeRF — CF-NeRF: Camera Parameter Free Neural Radiance Fields with Incremental Learning (AAAI 2024) · better on 1 of 1 shared result
- BayesRays — Bayes' Rays: Uncertainty Quantification for Neural Radiance Fields (CVPR 2024) · better on 1 of 1 shared result
- ActiveNeRF — ActiveNeRF: Learning Where to See with Uncertainty Estimation (ECCV 2022) · 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)
- Bayes' Rays: Uncertainty Quantification for Neural Radiance Fields (CVPR 2024 · similarity 0.89)
- NeRF On-the-go: Exploiting Uncertainty for Distractor-free NeRFs in the Wild (CVPR 2024 · similarity 0.88)
- ProvNeRF: Modeling per Point Provenance in NeRFs as a Stochastic Field (NeurIPS 2024 · similarity 0.88)
- NeRF-VPT: Learning Novel View Representations with Neural Radiance Fields via View Prompt Tuning (AAAI 2024 · similarity 0.87)
- Few-Shot Neural Radiance Fields under Unconstrained Illumination (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.
