Research map / 3D vision / Face and shape
Split-Layer: Enhancing Implicit Neural Representation by Maximizing the Dimensionality of Feature Space
Zhicheng Cai, Hao Zhu, Linsen Chen, Qiu Shen, 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 |
|---|---|---|---|
| replace hidden fully connected layers with split-layers (2 splits) | ReLU MLP · Image fitting (Natural) | PSNR (dB): 21.24 -> 30.89 | Improved |
| replace hidden layers with split-layers (2 splits) | PEMLP · Image fitting (Natural) | PSNR (dB): 29.60 -> 40.78 | Improved |
| replace hidden layers with split-layers (2 splits) | SIREN · Image fitting (Natural) | PSNR (dB): 38.52 -> 39.25 | Improved |
| replace hidden layers with split-layers (2 splits) | SIREN · CT reconstruction (x-ray colorectal) | PSNR (dB): 18.32 -> 29.11 | Improved |
| replace hidden layers with split-layers (2 splits) | ReLU MLP · CT reconstruction (x-ray colorectal) | PSNR (dB): 26.78 -> 30.19 | Improved |
| replace hidden layers with split-layers (2 splits) | PEMLP · 3D shape representation (Stanford) | Chamfer distance: 9.76e-6 -> 4.59e-6 | Improved |
| replace hidden layers with split-layers (2 splits) | ReLU MLP · 3D shape representation (Stanford) | Chamfer distance: 1.00e-4 -> 2.01e-5 | Improved |
| replace layers with split-layers | NeRF · NeRF synthetic (average) | PSNR (dB): 29.41 -> 30.20 | Improved |
| replace layers with split-layers | DVGO · NeRF synthetic (average) | PSNR (dB): 31.88 -> 32.10 | Improved |
| replace layers with split-layers | DINER · NeRF synthetic (average) | PSNR (dB): 32.09 -> 32.36 | Improved |
| add split-layer to DINER on Materials scene | DINER · NeRF synthetic (Materials) | PSNR (dB): 29.58 -> 29.59 | No clear effect |
| choose split number N near (0.17C)^(2/3) rather than other values (ablation) | split-layer INR · Image fitting with different widths | PSNR (dB) | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- FINER — FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024) · better on 2 of 2 shared results
- DINER — DINER: Depth-Aware Image-Based NEural Radiance Fields (CVPR 2023) · better on 1 of 1 shared result
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
- Learning Transferable Features for Implicit Neural Representations (NeurIPS 2024 · similarity 0.87)
- Implicit Representations via Operator Learning (ICML 2024 · similarity 0.87)
- Grids Often Outperform Implicit Neural Representation at Compressing Dense Signals (NeurIPS 2025 · similarity 0.87)
- VI^3NR: Variance Informed Initialization for Implicit Neural Representations (CVPR 2025 · similarity 0.87)
- MIRE: Matched Implicit Neural Representations (CVPR 2025 · similarity 0.85)
- Towards Lossless Implicit Neural Representation via Bit Plane Decomposition (CVPR 2025 · 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.
