Research map / 3D vision / NeRF and radiance fields
DeRF: Decomposed Radiance Fields
Check 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 |
|---|---|---|---|
| Increase the number of Voronoi heads at the same render cost | DeRF · Real Forward-Facing scenes | rendering quality | Helped |
| Use Voronoi decomposition rather than MLP-based decomposition (ablation) | DeRF · room | render time: 55.88s vs 32.89s; 76.51s vs 63.40s; 83.50s vs 127.32s | Mixed |
| Replace fixed grid cells with learned Voronoi cells (ablation) | DeRF · room | PSNR: 28.07 -> 28.55 | Helped |
| Replace fixed grid cells with learned Voronoi cells (ablation) | DeRF · room | SSIM: 0.88 -> 0.89 | Helped |
| Replace fixed grid cells with learned Voronoi cells (ablation) | DeRF · room | LPIPS: 0.28 -> 0.25 | Helped |
| Increase network capacity instead of decomposing the scene | NeRF · fern | inference time | Hurt |
| Train the decomposition before training the per-cell network heads | DeRF · training | training stability | Helped |
| Allow the uniformity loss to update the coarse density network (ablation) | DeRF pre-training · training | training stability | Hurt |
| Use a fixed grid instead of a learned decomposition (ablation) | DeRF · room | rendering quality: 28.55 vs 28.07 PSNR; 0.89 vs 0.88 SSIM; 0.25 vs 0.28 LPIPS | Hurt |
| Increase DeRF decomposition heads | DeRF · rendering quality | rendering quality | Mixed |
| Use a regular grid decomposition (ablation) | DeRF · real scenes | accuracy | Hurt |
Closest papers by meaning
- Switch-NeRF: Learning Scene Decomposition with Mixture of Experts for Large-scale Neural Radiance Fields (ICLR 2023 · similarity 0.90)
- NeRFReN: Neural Radiance Fields With Reflections (CVPR 2022 · similarity 0.90)
- HyRF: Hybrid Radiance Fields for Memory-efficient and High-quality Novel View Synthesis (NeurIPS 2025 · similarity 0.89)
- Rad-NeRF: Ray-decoupled Training of Neural Radiance Field (NeurIPS 2024 · similarity 0.89)
- NeRFPrior: Learning Neural Radiance Field as a Prior for Indoor Scene Reconstruction (CVPR 2025 · similarity 0.88)
- Mask-Based Modeling for Neural Radiance Fields (ICLR 2024 · 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.