Research map / 3D vision / Face and shape
Implicit neural representation design
Studies architectures, spectral bias and Fourier features for implicit neural representations.
52 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2017–2026) · an approach in Face and shape, keyword cluster “surface · mesh · implicit”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
13 papers in 2023–24 and 26 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 3.6% to 7.1%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2017 · Solid Harmonic Wavelet Scattering: Predicting Quantum Molecular Energy from Invariant Descriptors of 3D Electronic Densities (NeurIPS 2017 · 30 citations)
- 2019 · Fixing Implicit Derivatives: Trust-Region Based Learning of Continuous Energy Functions (NeurIPS 2019 · 2 citations)
- 2021 · Learning Continuous Image Representation With Local Implicit Image Function (CVPR 2021 · 770 citations)
- 2022 · CoordX: Accelerating Implicit Neural Representation with a Split MLP Architecture (ICLR 2022 · 8 citations)
- 2023 · Multiplicative Fourier Level of Detail (CVPR 2023 · 9 citations)
- 2024 · Learning Spatially Collaged Fourier Bases for Implicit Neural Representation (AAAI 2024 · 10 citations)
- 2025 · Adaptive Wavelet-Positional Encoding for High-Frequency Information Learning in Implicit Neural Representation (AAAI 2025 · 6 citations)
- 2026 · I-INR: Iterative Implicit Neural Representations (AAAI 2026 · 1 citation)
Earliest papers here
- Solid Harmonic Wavelet Scattering: Predicting Quantum Molecular Energy from Invariant Descriptors of 3D Electronic Densities (NeurIPS 2017 · 30 citations)
- Fixing Implicit Derivatives: Trust-Region Based Learning of Continuous Energy Functions (NeurIPS 2019 · 2 citations)
- Learning Continuous Image Representation With Local Implicit Image Function (CVPR 2021 · 770 citations)
What has been tried
35 papers here report 367 measured changes (a component added, removed or swapped, compared with a baseline): helped 66%, no clear effect 6%, hurt 22%, mixed 5%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Use learnable soft spatial masks to collage Fourier bases instead of globally combining bases | Kodak image fitting · PSNR | Helped | Learning Spatially Collaged Fourier Bases for Implicit Neural Representation (AAAI 2024 · 10 citations) | |
| Add Wavelet-Positional Encoding with HFP alongside global PE | LLFF · PSNR | +0.81 over PE (27.08 to 27.89) | Helped | Adaptive Wavelet-Positional Encoding for High-Frequency Information Learning in Implicit Neural Representation (AAAI 2025 · 6 citations) |
| Remove Gabor filters | CESM2 simulation · MSE | 4.073e-3 -> 1.758e-2 | Hurt | Continuous Field Reconstruction from Sparse Observations with Implicit Neural Networks (ICLR 2024 · 6 citations) |
| vary activation-parameter initialization across uniform and normal distributions | Parrot image representation · PSNR consistency | Helped | MIRE: Matched Implicit Neural Representations (CVPR 2025 · 5 citations) | |
| Add basic symmetric power transformation | DIV2K/Kodak, 1k iterations · PSNR | 30.30 -> 31.18 | Helped | Enhancing Implicit Neural Representations via Symmetric Power Transformation (AAAI 2025 · 3 citations) |
| Vary FINER's bias initialization range from U(−1/2, 1/2) to larger ranges | 2D image fitting · PSNR | Helped | FINER: Flexible Spectral-bias Tuning in Implicit NEural Representation by Variable-periodic Activation Functions (CVPR 2024 · 2 citations) | |
| condition the manager by concatenating manager and expert encoder outputs rather than no conditioning | Astronaut image · PSNR | 64.52 -> 81 .17 | Helped | Neural Experts: Mixture of Experts for Implicit Neural Representations (NeurIPS 2024 · 2 citations) |
| Set ReLU+NeFF dimensionality to L=16 rather than L=32 | Bach (7s) · SNR | 5.37 -> 24.76 | Helped | Representing Sounds as Neural Amplitude Fields: A Benchmark of Coordinate-MLPs and a Fourier Kolmogorov-Arnold Framework (AAAI 2025 · 1 citation) |
| Use SIREN rather than sigmoid activations | FRLL landmark alignment · landmark alignment MSE | NCF: 8.7E-3 -> 3.9E-5; NODE: 6.2E-2 -> 1.4E-4 | Helped | FLOWING: Implicit Neural Flows for Structure-Preserving Morphing (NeurIPS 2025 · 1 citation) |
| use interval arithmetic and IntervalNet instead of sampled points and an MLP target network | Airplane and Car mesh reconstruction · number of connected components | Helped | HyperCube: Implicit Field Representations of Voxelized 3D Models (Student Abstract) (AAAI 2024 · 1 citation) | |
| use learnable centers and learnable scaling instead of fixed patch locations | CIFAR-10-INRs · accuracy | 80.82±0.86% -> 81.33± 0.23% | Helped | Implicit Zoo: A Large-Scale Dataset of Neural Implicit Functions for 2D Images and 3D Scenes (NeurIPS 2024 · 1 citation) |
| Replace analytical activation statistics with Monte Carlo estimates | KODIM · PSNR | 79.73 -> 75.85 | Hurt | VI^3NR: Variance Informed Initialization for Implicit Neural Representations (CVPR 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 17 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- Kodak · 3 of 17 papers
- NeRF-Synthetic · 2 of 17 papers
- DIV2K · 2 of 17 papers
Baselines
- WIRE · 5 of 17 papers
- SIREN · 5 of 17 papers
- Gauss · 2 of 17 papers
- FINER · 2 of 17 papers
- BACON · 2 of 17 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 39 papers; from OpenAlex and the papers' affiliations).
- University of Edinburgh · 2 papers
- The University of Hong Kong · 2 papers
- Zhejiang University · 2 papers
- Shanghai Jiao Tong University · 2 papers
- Australian National University · 2 papers
- Google · 2 papers
- Amazon · 2 papers
- IMPA · 2 papers
- University of Coimbra · 2 papers
Use it
Download BibTeX of 32 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Neural implicit surface reconstruction (131 papers)
- Textured mesh generation (111 papers)
- Learned mesh and geometric representations (73 papers)
- Pixel-aligned implicit human and hand reconstruction (39 papers)
Elsewhere in the topic
- Gaussian and implicit head avatars (115 papers)
- Primitive-based shape abstraction (99 papers)
- Single-view 3D reconstruction (96 papers)
- Face reflectance and relightable rendering (85 papers)
- 3D morphable face model regression (80 papers)
- Full-body avatar reconstruction (78 papers)
- Articulated and non-rigid shape learning (75 papers)
- 3D shape generative modeling (74 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).