Research map / 3D vision / NeRF and radiance fields
Plenoxels: Radiance Fields Without Neural Networks
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 |
|---|---|---|---|
| Use trilinear rather than nearest-neighbor interpolation (ablation) | Plenoxels · NeRF synthetic scenes | PSNR: 27.17 -> 30.57 | Helped |
| Increase TV regularization when training with 25 rather than 100 views (ablation) | Plenoxels · NeRF synthetic scenes, 25 views | PSNR: 26.88 -> 28.25 | Helped |
| Increase grid resolution (ablation) | Plenoxels · NeRF synthetic scenes | PSNR: 23.49 -> 26.11 -> 28.46 -> 30.57 -> 31.71 | Helped |
| Use the Max/NeRF rendering formula instead of the Neural Volumes formula (ablation) | Plenoxels · NeRF synthetic scenes | PSNR: 27.54 -> 30.57 | Helped |
| Add sparsity regularization (ablation) | Plenoxels · NeRF forward-facing scenes | PSNR: 26.31 -> 26.29 | No clear effect |
| Add total variation regularization on opacity (ablation) | Plenoxels · NeRF forward-facing scenes | PSNR: 25.25 -> 26.29 | Helped |
| Add total variation regularization on spherical harmonics (ablation) | Plenoxels · NeRF forward-facing scenes | PSNR: 25.80 -> 26.29 | Helped |
| Use RMSProp rather than SGD for both opacity and spherical-harmonic coefficients (ablation) | Plenoxels · NeRF synthetic scenes | PSNR: 29.35 -> 30.57 | Helped |
| Use the delayed opacity learning-rate schedule rather than constant rates (ablation) | Plenoxels · NeRF synthetic scenes | PSNR: 30.13 -> 30.57 | Helped |
| Use tricubic rather than trilinear interpolation (ablation) | Plenoxels · NeRF synthetic scenes | Quality | No clear effect |
Closest papers by meaning
- Direct Voxel Grid Optimization: Super-Fast Convergence for Radiance Fields Reconstruction (CVPR 2022 · similarity 0.85)
- Sparse Voxels Rasterization: Real-time High-fidelity Radiance Field Rendering (CVPR 2025 · similarity 0.85)
- TinyNeRF: Towards 100 x Compression of Voxel Radiance Fields (AAAI 2023 · similarity 0.84)
- CVT-xRF: Contrastive In-Voxel Transformer for 3D Consistent Radiance Fields from Sparse Inputs (CVPR 2024 · similarity 0.84)
- Learning with Unreliability: Fast Few-shot Voxel Radiance Fields with Relative Geometric Consistency (CVPR 2024 · similarity 0.83)
- LAENeRF: Local Appearance Editing for Neural Radiance Fields (CVPR 2024 · similarity 0.82)
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.