Research map / 3D vision / NeRF and radiance fields
Coordinates Are NOT Lonely - Codebook Prior Helps Implicit Neural 3D representations
Fukun Yin, Wen Liu, Zilong Huang, Pei Cheng, Tao Chen, Gang YU
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 |
|---|---|---|---|
| replace MLP with codebook-prior CoCo network | VolSDF · H3DS sparse views 16-32 | PSNR: 23.922 -> 25.279 | Improved |
| replace MLP with codebook-prior CoCo network | VolSDF · BlendedMVS sparse views 16-32 | PSNR: 18.942 -> 19.594 | Improved |
| replace MLP with codebook-prior CoCo network | VolSDF · DTU sparse views 16-32 | PSNR: 26.609 -> 26.738 | Improved |
| replace MLP with codebook-prior CoCo network | VolSDF · H3DS few views 5-8 | PSNR: 18.939 -> 19.084 | No clear effect |
| replace MLP with codebook-prior CoCo network | VolSDF · DTU few views 5-8 | PSNR: 18.692 -> 19.624 | Improved |
| replace MLP with codebook-prior CoCo network | mask-based VolSDF · DTU 3 views | LPIPS: 0.491 -> 0.396 | Improved |
| replace MLP with codebook-prior CoCo network | mask-based VolSDF · DTU 3 views | PSNR: 13.929 -> 14.761 | Improved |
| use per-scene learnable codebook instead of fixed pre-trained codebook (ablation) | CoCo-INR · DTU scans 24/37/40 few views | PSNR: 15.622 -> 13.315 | Hurt |
| use per-scene learnable codebook instead of fixed pre-trained codebook (ablation) | CoCo-INR · DTU scans 24/37/40 few views | LPIPS: 0.283 -> 0.298 | Hurt |
| reduce codebook items 16384 to 8192 (ablation) | CoCo-INR · DTU scans 24/37/40 few views | PSNR: 15.622 -> 15.463 | Hurt |
| reduce codebook items 16384 to 4096 (ablation) | CoCo-INR · DTU scans 24/37/40 few views | PSNR: 15.622 -> 14.042 | Hurt |
| reduce codebook items 16384 to 2048 (ablation) | CoCo-INR · DTU scans 24/37/40 few views | PSNR: 15.622 -> 14.762 | Hurt |
Closest papers by meaning
- H3D-Net: Few-Shot High-Fidelity 3D Head Reconstruction (ICCV 2021 · similarity 0.88)
- Coordinate Quantized Neural Implicit Representations for Multi-view Reconstruction (ICCV 2023 · similarity 0.88)
- Implicit Shape and Appearance Priors for Few-Shot Full Head Reconstruction (TPAMI 2025 · similarity 0.87)
- Surface Normal Clustering for Implicit Representation of Manhattan Scenes (ICCV 2023 · similarity 0.87)
- 3D Reconstruction with Generalizable Neural Fields using Scene Priors (ICLR 2024 · similarity 0.85)
- PM-INR: Prior-Rich Multi-Modal Implicit Large-Scale Scene Neural Representation (AAAI 2024 · similarity 0.85)
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.
