Research map / 3D vision / NeRF and radiance fields
NeRF for scene decomposition and understanding
Uses radiance fields for geometry decomposition, object selection, segmentation and detection.
30 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2022–2026) · an approach in NeRF and radiance fields, keyword cluster “radiance fields · fields nerf · nerfs”
State History Tried Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
17 papers in 2023–24 and 10 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 6.4% to 11.1%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2022 · BACON: Band-Limited Coordinate Networks for Multiscale Scene Representation (CVPR 2022 · 119 citations)
- 2023 · Neural Fourier Filter Bank (CVPR 2023 · 27 citations)
- 2024 · Improving Physics-Augmented Continuum Neural Radiance Field-Based Geometry-Agnostic System Identification with Lagrangian Particle Optimization (CVPR 2024 · 8 citations)
- 2025 · GO-N3RDet: Geometry Optimized NeRF-enhanced 3D Object Detector (CVPR 2025 · 5 citations)
- 2026 · Dynamic Novel View Synthesis in High Dynamic Range (ICLR 2026 · 0 citations)
Earliest papers here
- BACON: Band-Limited Coordinate Networks for Multiscale Scene Representation (CVPR 2022 · 119 citations)
- Neural Volumetric Object Selection (CVPR 2022 · 49 citations)
- Virtual Elastic Objects (CVPR 2022 · 20 citations)
What has been tried
25 papers here report 263 measured changes (a component added, removed or swapped, compared with a baseline): helped 70%, no clear effect 3%, hurt 21%, mixed 6%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Optimize particle features only instead of both features and positions | Sparse-view test views · PSNR | 29.22 -> 28.22 | Hurt | Improving Physics-Augmented Continuum Neural Radiance Field-Based Geometry-Agnostic System Identification with Lagrangian Particle Optimization (CVPR 2024 · 8 citations) |
| Add blending-weight regularization Lb | Videos · L2 | 0.0322 -> 0.0339 | Hurt | In-N-Out: Faithful 3D GAN Inversion with Volumetric Decomposition for Face Editing (CVPR 2024 · 8 citations) |
| add PEOM to naive voxel feature construction | ScanNet · mAP@0.25 | 48.4 -> 55.2 | Helped | GO-N3RDet: Geometry Optimized NeRF-enhanced 3D Object Detector (CVPR 2025 · 5 citations) |
| Use gated image formation with separate ambient and reflection components instead of a single end-to-end intensity field | Test Data – Night · MAE [m] | 19.95 -> 8.35 | Helped | Gated Fields: Learning Scene Reconstruction from Gated Videos (CVPR 2024 · 3 citations) |
| add geometric prior | ScanNet surface reconstruction · Acc | 0.079 -> 0.046 | Helped | 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 2024 · 2 citations) |
| Set Lout equal to Lin for the outer scene | Mip-NeRF 360, k=10 · PSNR | 24.71 -> 26.27 | Helped | Alpha Invariance: On Inverse Scaling Between Distance and Volume Density in Neural Radiance Fields (CVPR 2024 · 1 citation) |
| add normal priors | ScanNet (12 scenes) · F-score | 0.353 -> 0.766 | Helped | H2O-SDF: Two-phase Learning for 3D Indoor Reconstruction using Object Surface Fields (ICLR 2024 · 1 citation) |
| Add negative motion fusion | EPIC Fields UDOS · mAP Dyn | 55.58 -> 63.73 | Helped | Layered Motion Fusion: Lifting Motion Segmentation to 3D in Egocentric Videos (CVPR 2025 · 0 citations) |
| add mass loss | SfC dataset, cavity size average · CD (×103↓) | 0.503 -> 0.165 | Helped | Structure from Collision (CVPR 2025 · 0 citations) |
| Replace Transformer with flattened-input MLP | ModelNet40 (UDF) · accuracy | 87.0 -> 41.6 | Hurt | Neural Processing of Tri-Plane Hybrid Neural Fields (ICLR 2024 · 0 citations) |
| increase the perturbation bound from ϵ=1.0 to ϵ=5.0 | IM3D, severity 1–5 · average accuracy | 0.6857 -> 0.8513 | Helped | You Always Recognize Me (YARM): Robust Texture Synthesis Against Multi-View Corruption (ICML 2025) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 27 papers; from OpenAlex and the papers' affiliations).
- NTT Corporation · 2 papers
- University of Maryland, College Park · 2 papers
Use it
Download BibTeX of 30 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Sparse-input NeRF regularization (106 papers)
- Dynamic and editable NeRF (88 papers)
- Single-view and streaming radiance fields (49 papers)
Elsewhere in the topic
- Inverse rendering with neural fields (110 papers)
- Few-shot neural rendering architectures (53 papers)
- 3D-aware generative radiance fields (20 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).