Research map / 3D vision / NeRF and radiance fields
TensoRF: Tensorial Radiance Fields
Anpei Chen, Zexiang Xu, Andreas Geiger, Jingyi Yu, Hao Su
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 L1 regularization (ablation) | TensoRF VM-192 · Synthetic-NeRF | PSNR: 32.43 -> 33.14 | Improved |
| add L1 regularization (ablation) | TensoRF CP-384 · Synthetic-NeRF | PSNR: 31.23 -> 31.56 | Improved |
| add L1 regularization (ablation) | TensoRF VM-48 · Synthetic-NeRF | PSNR: 31.71 -> 32.39 | Improved |
| use VM decomposition instead of CP (smallest VM vs best CP) (ablation) | TensoRF · Synthetic-NeRF | PSNR: 31.56 -> 31.81 | Improved |
| double VM components from 192 to 384 (ablation) | TensoRF-VM · Synthetic-NeRF (300^3) | PSNR: 33.14 -> 33.21 | No clear effect |
| raise grid resolution from 300^3 to 500^3 (ablation) | TensoRF-VM-192 · Synthetic-NeRF | PSNR: 33.14 -> 33.31 | Improved |
| increase CP components from 48 to 384 (ablation) | TensoRF-CP (200^3) · Synthetic-NeRF | PSNR: 27.98 -> 30.47 | Improved |
| train longer from 15k to 60k steps (ablation) | TensoRF VM-192 · Synthetic-NeRF | PSNR: 32.52 -> 33.44 | Improved |
| train longer from 5k to 60k steps (ablation) | TensoRF CP-384 · Synthetic-NeRF | PSNR: 28.37 -> 32.03 | Improved |
| use VM-192 without spherical harmonics (MLP decoder) instead of SH shading (ablation) | TensoRF VM-192 · Synthetic-NeRF | PSNR: 32.00 -> 33.14 | Improved |
| replace VM decomposition with CP at 384 components, 500^3 (ablation) | TensoRF · Synthetic-NeRF | model size and PSNR: 31.56 PSNR at 3.9 MB vs 33.39 at 397 MB | Mixed |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Plenoxels — Plenoxels: Radiance Fields Without Neural Networks (CVPR 2022) · better on 2 of 2 shared results
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
- TensoIR: Tensorial Inverse Rendering (CVPR 2023 · similarity 0.92)
- Direct Voxel Grid Optimization: Super-Fast Convergence for Radiance Fields Reconstruction (CVPR 2022 · similarity 0.91)
- Sparse Voxels Rasterization: Real-time High-fidelity Radiance Field Rendering (CVPR 2025 · similarity 0.90)
- HyRF: Hybrid Radiance Fields for Memory-efficient and High-quality Novel View Synthesis (NeurIPS 2025 · similarity 0.89)
- Improving Robustness for Joint Optimization of Camera Pose and Decomposed Low-Rank Tensorial Radiance Fields (AAAI 2024 · similarity 0.88)
- Learning with Unreliability: Fast Few-shot Voxel Radiance Fields with Relative Geometric Consistency (CVPR 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.
