Research map / Image generation / Image restoration
Raw and low-light noise modeling
Model camera noise physics to train raw denoisers.
53 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Image restoration, keyword cluster “image denoising · blur · noise”
State History Benchmarks Tried Preprints Who Export
Where it stands
Few recent reports The tracked benchmarks have at most one report from 2025–26 that could beat the old record, too few to say whether they are still moving.
18 papers in 2023–24 and 14 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 5.8% to 3.2%.
No longer reported by papers here: SIDD Benchmark (SSIM, last in 2024).
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
- 2016 · A Holistic Approach to Cross-Channel Image Noise Modeling and Its Application to Image Denoising (CVPR 2016 · 290 citations)
- 2017 · Benchmarking Denoising Algorithms With Real Photographs (CVPR 2017 · 689 citations)
- 2018 · A High-Quality Denoising Dataset for Smartphone Cameras (CVPR 2018 · 898 citations)
- 2019 · A Poisson-Gaussian Denoising Dataset With Real Fluorescence Microscopy Images (CVPR 2019 · 204 citations)
- 2020 · Blindly Assess Image Quality in the Wild Guided by a Self-Adaptive Hyper Network (CVPR 2020 · 840 citations)
- 2021 · FBI-Denoiser: Fast Blind Image Denoiser for Poisson-Gaussian Noise (CVPR 2021 · 58 citations)
- 2022 · Model-Based Image Signal Processors via Learnable Dictionaries (AAAI 2022 · 40 citations)
- 2023 · Spatially Adaptive Self-Supervised Learning for Real-World Image Denoising (CVPR 2023 · 59 citations)
- 2024 · Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · 18 citations)
- 2025 · Complementary Advantages: Exploiting Cross-Field Frequency Correlation for NIR-Assisted Image Denoising (CVPR 2025 · 11 citations)
- 2026 · Real Noise Decoupling for Hyperspectral Image Denoising (AAAI 2026 · 2 citations)
Earliest papers here
- A Holistic Approach to Cross-Channel Image Noise Modeling and Its Application to Image Denoising (CVPR 2016 · 290 citations)
- From Noise Modeling to Blind Image Denoising (CVPR 2016 · 87 citations)
- Benchmarking Denoising Algorithms With Real Photographs (CVPR 2017 · 689 citations)
Benchmark records
The best number each year that beat every earlier one, among papers of this approach reporting their own method. Numbers are extracted by AI from each paper's text; check the paper before citing.
SIDD Benchmark · SSIM (higher is better)
5 papers, 2022–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2022 | 0.912 | Modeling sRGB Camera Noise With Normalizing Flows (CVPR 2022 · 25 citations) |
| 2023 | 0.934 | Spatially Adaptive Self-Supervised Learning for Real-World Image Denoising (CVPR 2023 · 59 citations) |
| 2024 | 0.935 | sRGB Real Noise Modeling via Noise-Aware Sampling with Normalizing Flows (ICLR 2024) |
SIDD benchmark · PSNR (dB) (higher is better)
4 papers, 2023–2025 · one report since 2024, not better
| Year | Best so far | Set by |
|---|---|---|
| 2023 | 36.82 | sRGB Real Noise Synthesizing With Neighboring Correlation-Aware Noise Model (CVPR 2023 · 15 citations) |
| 2024 | 37.69 | Pseudo-Siamese Blind-spot Transformers for Self-Supervised Real-World Denoising (NeurIPS 2024 · 2 citations) |
What has been tried
41 papers here report 390 measured changes (a component added, removed or swapped, compared with a baseline): helped 79%, no clear effect 4%, hurt 14%, mixed 2%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add adaptive masking | SIDD · PSNR (dB) | 33.12 -> 34.45 | Helped | Masked and Shuffled Blind Spot Denoising for Real-World Images (CVPR 2024 · 18 citations) |
| replace SNN neurons with ANN ReLU | LED · Denoising accuracy | 86.0 -> 84.5 | Hurt | LED: A Large-scale Real-world Paired Dataset for Event Camera Denoising (CVPR 2024 · 17 citations) |
| Add FDSM to the two-stage U-Net baseline | DVD test, σ = 4 · PSNR | 29.13 -> 29.64 | Helped | Complementary Advantages: Exploiting Cross-Field Frequency Correlation for NIR-Assisted Image Denoising (CVPR 2025 · 11 citations) |
| replace calibrated K with hypothesized K on SID and ELD | SID and ELD · PSNR | 40.90 / 0.9487 vs 40.85 / 0.9478; 46.37 / 0.9832 vs 46.43 / 0.9834 | No clear effect | Noise Modeling in One Hour: Minimizing Preparation Efforts for Self-supervised Low-Light RAW Image Denoising (CVPR 2025 · 8 citations) |
| Lower masking probability from p0=0.50 to p0=0.10 | SIDD-Validation · PSNR | 34.84 -> 35.31 | Helped | Zero-Shot Blind-spot Image Denoising via Implicit Neural Sampling (CVPR 2025 · 3 citations) |
| Add the N2N loss | 70S ribosome · F1 score | 0.786 -> 0.810 | Helped | DRACO: A Denoising-Reconstruction Autoencoder for Cryo-EM (NeurIPS 2024 · 3 citations) |
| add the PDA module | LIVEC · PLCC | 0.881 -> 0.911 | Helped | Feature Denoising Diffusion Model for Blind Image Quality Assessment (AAAI 2025 · 2 citations) |
| Add directional self-attention and the four-branch structure instead of grid SA with one branch | SIDD-Validation · PSNR (dB) | 23.96 -> 37.63 | Helped | Pseudo-Siamese Blind-spot Transformers for Self-Supervised Real-World Denoising (NeurIPS 2024 · 2 citations) |
| Add pixel SSN loss to denoising loss | BSD500 test · PSNR | 32.77 -> 32.78 | Helped | Soft Superpixel Neighborhood Attention (NeurIPS 2024 · 2 citations) |
| Use noise-aware sampling over randomly selected latent distributions | SIDD-Validation · KLD | 0.4700 -> 0.0291 | Helped | sRGB Real Noise Modeling via Noise-Aware Sampling with Normalizing Flows (ICLR 2024) |
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 32 papers; from OpenAlex and the papers' affiliations).
- National University of Singapore · 3 papers
- Tsinghua University · 2 papers
- Beijing Institute of Technology · 2 papers
- Purdue University · 2 papers
- Sony · 2 papers
Use it
Download BibTeX of 39 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
- Image deblurring (73 papers)
- Self-supervised image denoising (60 papers)
- Diffusion priors for denoising and deblurring (41 papers)
Elsewhere in the topic
- Dehazing and deraining (118 papers)
- Efficient restoration architectures (113 papers)
- Inverse imaging with learned priors (109 papers)
- Image harmonization and color transformation (102 papers)
- Reflection removal (95 papers)
- Pansharpening and spectral fusion (84 papers)
- Low-light image enhancement (72 papers)
- Learned image compression (69 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).