Research map / 3D vision / Depth and stereo
Learned stereo with specialized sensors
Estimates depth from stereo pairs with active, polarization, normal or exposure cues.
53 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 Checklist Preprints Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
5 papers in 2023–24 and 19 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 1.6% to 4.6%.
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 · HyperDepth: Learning Depth From Structured Light Without Matching (CVPR 2016 · 108 citations)
- 2017 · CATS: A Color and Thermal Stereo Benchmark (CVPR 2017 · 70 citations)
- 2018 · Left-Right Comparative Recurrent Model for Stereo Matching (CVPR 2018 · 83 citations)
- 2019 · DAVANet: Stereo Deblurring With View Aggregation (CVPR 2019 · 103 citations)
- 2020 · Adaptive Unimodal Cost Volume Filtering for Deep Stereo Matching (AAAI 2020 · 179 citations)
- 2021 · Deep Event Stereo Leveraged by Event-to-Image Translation (AAAI 2021 · 31 citations)
- 2022 · Depth Estimation by Combining Binocular Stereo and Monocular Structured-Light (CVPR 2022 · 14 citations)
- 2023 · BEVStereo: Enhancing Depth Estimation in Multi-View 3D Object Detection with Temporal Stereo (AAAI 2023 · 216 citations)
- 2024 · EventPS: Real-Time Photometric Stereo Using an Event Camera (CVPR 2024 · 22 citations)
- 2025 · DEFOM-Stereo: Depth Foundation Model Based Stereo Matching (CVPR 2025 · 43 citations)
- 2026 · GMT: Effective Global Framework for Multi-Camera Multi-Target Tracking (CVPR 2026 · 2 citations)
Earliest papers here
- HyperDepth: Learning Depth From Structured Light Without Matching (CVPR 2016 · 108 citations)
- Stereo Matching With Color and Monochrome Cameras in Low-Light Conditions (CVPR 2016 · 46 citations)
- Using Self-Contradiction to Learn Confidence Measures in Stereo Vision (CVPR 2016 · 32 citations)
What has been tried
41 papers here report 367 measured changes (a component added, removed or swapped, compared with a baseline): helped 81%, no clear effect 2%, hurt 15%, 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 |
|---|---|---|---|---|
| increase event-triggering noise std from 0.05 to 0.2 | DiLiGenT-Ev · MAE | 23.6 -> 28.1 | Hurt | EventPS: Real-Time Photometric Stereo Using an Event Camera (CVPR 2024 · 22 citations) |
| use ViT-B instead of ViT Small | RMVDB average · rel | 3.40 -> 3.00 | Helped | MVSAnywhere: Zero-Shot Multi-View Stereo (CVPR 2025 · 11 citations) |
| Replace convolutional decoding with INMF | SceneFlow test · EPE (px) | 0.88 -> 0.83 | Helped | Any-Stereo: Arbitrary Scale Disparity Estimation for Iterative Stereo Matching (AAAI 2024 · 7 citations) |
| add vertical optical-flow self-supervision | Semi-Truck Highway · MAE | 0.023 -> 0.015 | Helped | Flow-Guided Online Stereo Rectification for Wide Baseline Stereo (CVPR 2024 · 6 citations) |
| Add MobileNet blocks while retaining concatenation cost volume | Active-Event-Stereo test · Runtime (ms) | 33.8 -> 6.2 | Helped | Active Event-based Stereo Vision (CVPR 2025 · 5 citations) |
| Replace per-view SVD-based UPS with HO-GSVD | BUNNY, 4 views · MAE (UPS) | 0.62 -> 0.13 | Helped | MVCPS-NeuS: Multi-view Constrained Photometric Stereo for Neural Surface Reconstruction (CVPR 2024 · 5 citations) |
| Add illumination-invariant, sub-pixel, and fusion modules to Base | QPD2K · epe | 0.353 -> 0.201 | Helped | All-directional Disparity Estimation for Real-world QPD Images (CVPR 2025 · 1 citation) |
| Replace RAFT-Stereo [24] with backbone [51] | Ablation study · disparity MAE [px] | 1.76 -> 1.46 | Helped | Dual Exposure Stereo for Extended Dynamic Range 3D Imaging (CVPR 2025 · 0 citations) |
| replace pixel-wise DOE encoding with rank-2 encoding | Scene Flow testset, 0.8 m⁻¹ · PSNR | 31.88 -> 32.96 | Helped | Learned Binocular-Encoding Optics for RGBD Imaging Using Joint Stereo and Focus Cues (CVPR 2025 · 0 citations) |
| Use multi-pixel output instead of single-pixel output | MVSEC indoor_flying, stereo + F_orig · #Points [million] | 0.79 -> 4.32 | Helped | DERD-Net: Learning Depth from Event-based Ray Densities (NeurIPS 2025 · 0 citations) |
What reviewers will expect
Datasets and baselines named in the 10 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- Tartan Air · 2 of 10 papers
- InStereo2k · 2 of 10 papers
Baselines
- SGM · 2 of 10 papers
- DDES · 2 of 10 papers
- DLNR · 2 of 10 papers
- RAFT-Stereo · 2 of 10 papers
Recent preprints
Loading recent arXiv preprints…
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
- Cost-volume multi-view stereo (107 papers)
- Photometric stereo (43 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).