Research map / 3D vision / Depth and stereo
Photometric stereo
Estimates surface normals from varying illumination, including uncalibrated and non-Lambertian cases.
43 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Depth and stereo, keyword cluster “view stereo · photometric stereo · stereo matching”
State History Tried Baselines 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.
10 papers in 2023–24 and 7 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic shrank, from 3.3% to 1.7%.
Open questions the data points to
- Papers report PS-NeRF on DiLiGenT-MV, Buddha (Chamfer distance) anywhere from 0.28 to 0.314: comparisons against it depend on whose number you take.
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 Benchmark Dataset and Evaluation for Non-Lambertian and Uncalibrated Photometric Stereo (CVPR 2016 · 161 citations)
- 2017 · Polarimetric Multi-View Stereo (CVPR 2017 · 142 citations)
- 2018 · Neural Inverse Rendering for General Reflectance Photometric Stereo (ICML 2018 · 63 citations)
- 2019 · InverseRenderNet: Learning Single Image Inverse Rendering (CVPR 2019 · 161 citations)
- 2020 · Lighthouse: Predicting Lighting Volumes for Spatially-Coherent Illumination (CVPR 2020 · 103 citations)
- 2021 · Uncalibrated Neural Inverse Rendering for Photometric Stereo of General Surfaces (CVPR 2021 · 50 citations)
- 2022 · Uncertainty-Aware Deep Multi-View Photometric Stereo (CVPR 2022 · 32 citations)
- 2023 · Scalable, Detailed and Mask-Free Universal Photometric Stereo (CVPR 2023 · 52 citations)
- 2024 · NeRSP: Neural 3D Reconstruction for Reflective Objects with Sparse Polarized Images (CVPR 2024 · 13 citations)
- 2025 · EventPSR: Surface Normal and Reflectance Estimation from Photometric Stereo Using an Event Camera (CVPR 2025 · 3 citations)
- 2026 · ReLi3D: Relightable Multi-view 3D Reconstruction with Disentangled Illumination (ICLR 2026 · 0 citations)
Earliest papers here
- A Benchmark Dataset and Evaluation for Non-Lambertian and Uncalibrated Photometric Stereo (CVPR 2016 · 161 citations)
- Unbiased Photometric Stereo for Colored Surfaces: A Variational Approach (CVPR 2016 · 39 citations)
- Uncalibrated Photometric Stereo by Stepwise Optimization Using Principal Components of Isotropic BRDFs (CVPR 2016 · 6 citations)
What has been tried
39 papers here report 282 measured changes (a component added, removed or swapped, compared with a baseline): helped 73%, no clear effect 4%, hurt 18%, 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 |
|---|---|---|---|---|
| add the photometric loss Lp | SMVP3D DRAGON · Chamfer distance | 5.17 -> 4.77 | Helped | NeRSP: Neural 3D Reconstruction for Reflective Objects with Sparse Polarized Images (CVPR 2024 · 13 citations) |
| add reflectance maps to the normal-based reconstruction | DiLiGenT-MV average · Chamfer distance | 0.26 -> 0.23 | Helped | RNb-NeuS: Reflectance and Normal-based Multi-View 3D Reconstruction (CVPR 2024 · 12 citations) |
| add boundary-based light initialization instead of random initialization | Synthetic dataset · PU-PSNR | 16.86 -> 22.06 | Helped | Spin-UP: Spin Light for Natural Light Uncalibrated Photometric Stereo (CVPR 2024 · 7 citations) |
| Initialize learned patterns with flat gray instead of heuristic patterns | 3D-printed test dataset · reconstruction error | 0.3930 -> 0.0466 | Helped | Differentiable Display Photometric Stereo (CVPR 2024 · 5 citations) |
| use optimal diffuse albedo instead of an MLP diffuse-albedo network | two viewpoints, wide baseline · 3D reconstruction quality | Helped | Sparse Views Near Light: A Practical Paradigm for Uncalibrated Point-light Photometric Stereo (CVPR 2024 · 4 citations) | |
| Add vertex alignment to decimation | DiLiGenT-MV, 1k vertices · RMSE (mm) | 1.64 -> 1.59 | Helped | Feature-Preserving Mesh Decimation for Normal Integration (CVPR 2025 · 2 citations) |
| Add shape-based masks to exclude distorted event-profile regions | Glossy, Pole, and MulRefs · MAE | Helped | PS-EIP: Robust Photometric Stereo Based on Event Interval Profile (CVPR 2025 · 2 citations) | |
| Add the FNIN integration network rather than using BiNI (k=0) | DiLiGenT (10 objects) · MAE (mm) | 4.59 -> 4.23 | Helped | FNIN: A Fourier Neural Operator-based Numerical Integration Network for Surface-from-gradients (AAAI 2025 · 1 citation) |
Baseline numbers that disagree
The same baseline on the same benchmark and metric, reported more than 5% apart by different papers here. Training budgets, data splits and reimplementations differ, so state which number you compare with.
| Baseline | Benchmark | Metric | Reported | Papers |
|---|---|---|---|---|
| PS-NeRF | DiLiGenT-MV, Buddha | Chamfer distance | 0.28 – 0.314 | 2 papers, e.g. CVPR 2024 and CVPR 2023 |
What reviewers will expect
Datasets and baselines named in the 10 papers of this approach from 2024–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- DiLiGenT-MV · 3 of 10 papers
- DiLiGenT · 2 of 10 papers
- LUCES · 2 of 10 papers
Baselines
- SDM-UniPS · 4 of 10 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 17 papers; from OpenAlex and the papers' affiliations).
- Peking University · 3 papers
- Osaka University · 3 papers
- National Institute of Informatics · 2 papers
- Beijing University of Posts and Telecommunications · 2 papers
Use it
Download BibTeX of 33 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
- Cost-volume multi-view stereo (107 papers)
- Learned stereo with specialized sensors (53 papers)
Elsewhere in the topic
- Feed-forward 3D reconstruction (133 papers)
- Self-supervised monocular depth (116 papers)
- Scene coordinate regression for localization (112 papers)
- Learned optical and scene flow (108 papers)
- Neural light field and illumination imaging (107 papers)
- Camera model and calibration for SfM (99 papers)
- Non-line-of-sight imaging (93 papers)
- Dense depth prediction and completion (85 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).