Research map / Label-efficient and robust learning / Recognition and long-tail
Polynomial Neural Fields for Subband Decomposition and Manipulation
Guandao Yang, Sagie Benaim, Varun Jampani, Kyle Genova, Jonathan T. Barron, Thomas Funkhouser, Bharath Hariharan, Serge Belongie
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 BACON with PNF | neural field · DIV2K image fitting | PSNR: 28.67 -> 29.47 | Improved |
| replace SIREN with PNF | neural field · DIV2K image fitting | PSNR: 29.22 -> 29.47 | Improved |
| replace BACON-L with PNF | neural field · DIV2K image fitting | PSNR: 29.44 -> 29.47 | No clear effect |
| replace BACON with PNF | neural field · 3D shape fitting | Chamfer Distance (x10^6): 2.60 -> 2.25 | Improved |
| replace SIREN with PNF | neural field · 3D shape fitting | Chamfer Distance (x10^6): 9.00 -> 2.25 | Improved |
| replace BACON with PNF | neural field · 3D shape fitting | F-score: 99.84% -> 99.97% | Improved |
| replace BACON with PNF at 300 epochs | neural field · NeRF fitting 64^2, 1x scale | PSNR: 28.07 -> 29.89 | Improved |
| replace BACON with PNF at 300 epochs | neural field · NeRF fitting 64^2, average | SSIM: 0.936 -> 0.934 | Hurt |
| replace BACON with PNF at 500 epochs | neural field · NeRF fitting 64^2, 1/8x scale | PSNR: 32.10 -> 31.37 | Hurt |
| train 300 to 500 epochs | PNF · NeRF fitting 64^2, average | PSNR: 31.29 -> 30.84 | Hurt |
| replace BACON with PNF | neural field · cameraman image fitting | time to 36 PSNR (s): 177 -> 96 | Improved |
| replace BACON with PNF | neural field · cameraman image fitting | time per step (s): 0.16 -> 0.64 | Hurt |
Closest papers by meaning
- BANF: Band-Limited Neural Fields for Levels of Detail Reconstruction (CVPR 2024 · similarity 0.88)
- Neural Fields as Distributions: Signal Processing Beyond Euclidean Space (CVPR 2024 · similarity 0.87)
- FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024 · similarity 0.85)
- Coordinate-Aware Modulation for Neural Fields (ICLR 2024 · similarity 0.85)
- Beyond Regular Grids: Fourier-Based Neural Operators on Arbitrary Domains (ICML 2024 · similarity 0.85)
- Learning Spatially Collaged Fourier Bases for Implicit Neural Representation (AAAI 2024 · similarity 0.84)
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.
