Research map / 3D vision / Face and shape
Disorder-Invariant Implicit Neural Representation
Hao Zhu, Shaowen Xie, Zhen Liu, Fengyi Liu, Qi Zhang, You Zhou, Yi Lin, Zhan Ma, Xun Cao
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 hash-table to SIREN backbone (ablation) | SIREN · 2D image fitting | PSNR: increased 30dB | Improved |
| add hash-table to PE+MLP/MLP backbone (ablation) | MLP · 2D image fitting | PSNR: increased 17dB | Improved |
| DINER (MLP) versus InstantNGP | DINER MLP · 2D image fitting | PSNR: advantage of 6dB | Improved |
| increase hash-table width up to number of image channels (ablation) | DINER · 2D image fitting 1/2/3-channel | PSNR | Improved |
| increase hash-table width beyond number of channels (ablation) | DINER · 2D image fitting | PSNR | No clear effect |
| increase hash-table width up to rank of linearly dependent channels (ablation) | DINER · Multispectral 6- and 7-channel images | PSNR | Improved |
| use SIREN backbone instead of MLP in DINER at 10000 epochs (ablation) | DINER · 2D image fitting | PSNR: 59.64dB vs 59.53dB | No clear effect |
| DINER instead of NeRV on video representation | DINER · UVG video fitting | PSNR: 36.08 dB -> 50.74 dB | Improved |
| DINER instead of SIREN on video representation (ablation) | SIREN 4x64 · UVG video fitting | PSNR: 21.22 dB -> 50.74 dB | Improved |
| add hash-table to SIREN on video (ablation) | SIREN · UVG video fitting | training time: 706s -> 1309s | Hurt |
| DINER instead of DeCAF for refractive index recovery | DINER · Granulocyte phantom | training time: 401 minutes -> 91 minutes | Improved |
| DINER instead of PhysenNet in lensless imaging | DINER · Real lensless imaging | PSNR: >7 dB | Improved |
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 3 of 4 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
- Grids Often Outperform Implicit Neural Representation at Compressing Dense Signals (NeurIPS 2025 · similarity 0.87)
- FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024 · similarity 0.87)
- Implicit Representations via Operator Learning (ICML 2024 · similarity 0.87)
- MIRE: Matched Implicit Neural Representations (CVPR 2025 · similarity 0.86)
- Towards Lossless Implicit Neural Representation via Bit Plane Decomposition (CVPR 2025 · similarity 0.86)
- Learning Transferable Features for Implicit Neural Representations (NeurIPS 2024 · similarity 0.86)
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.
