Research map / 3D vision / Human pose estimation
Learned camera pose estimation
Estimates absolute and relative camera pose using learned features, pose regression and geometric solvers.
96 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2016–2026) · an approach in Human pose estimation, keyword cluster “object pose · 6d · hand”
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.
25 papers in 2023–24 and 31 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 9.2% to 10.3%.
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 · 6D Dynamic Camera Relocalization From Single Reference Image (CVPR 2016 · 17 citations)
- 2017 · Simultaneous Facial Landmark Detection, Pose and Deformation Estimation Under Facial Occlusion (CVPR 2017 · 79 citations)
- 2018 · Disentangling 3D Pose in a Dendritic CNN for Unconstrained 2D Face Alignment (CVPR 2018 · 168 citations)
- 2019 · Understanding the Limitations of CNN-Based Absolute Camera Pose Regression (CVPR 2019 · 408 citations)
- 2020 · Deep Global Registration (CVPR 2020 · 559 citations)
- 2021 · Back to the Feature: Learning Robust Camera Localization From Pixels To Pose (CVPR 2021 · 295 citations)
- 2022 · EPro-PnP: Generalized End-to-End Probabilistic Perspective-N-Points for Monocular Object Pose Estimation (CVPR 2022 · 189 citations)
- 2023 · SliceMatch: Geometry-Guided Aggregation for Cross-View Pose Estimation (CVPR 2023 · 40 citations)
- 2024 · Neural Refinement for Absolute Pose Regression with Feature Synthesis (CVPR 2024 · 26 citations)
- 2025 · Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (CVPR 2025 · 31 citations)
- 2026 · Globally Optimal Pose from Orthographic Silhouettes (CVPR 2026 · 0 citations)
Earliest papers here
- 6D Dynamic Camera Relocalization From Single Reference Image (CVPR 2016 · 17 citations)
- Simultaneous Facial Landmark Detection, Pose and Deformation Estimation Under Facial Occlusion (CVPR 2017 · 79 citations)
- Disentangling 3D Pose in a Dendritic CNN for Unconstrained 2D Face Alignment (CVPR 2018 · 168 citations)
What has been tried
74 papers here report 772 measured changes (a component added, removed or swapped, compared with a baseline): helped 78%, no clear effect 3%, hurt 16%, mixed 3%. 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 fully symmetric shared-weight architecture instead of asymmetric branches | ScanNet1500 · AUC@5 | 32.71 -> 34.79 | Helped | Reloc3r: Large-Scale Training of Relative Camera Pose Regression for Generalizable, Fast, and Accurate Visual Localization (CVPR 2025 · 31 citations) |
| Use camera-pose estimates as geometry-learning conditioning | BlendedMVS test set · Overall Chamfer distance | 0.0319 -> 0.0288 | Helped | FLARE: Feed-forward Geometry, Appearance and Camera Estimation from Uncalibrated Sparse Views (CVPR 2025 · 28 citations) |
| Remove exposure-adaptive ACT | Cambridge Shop Facade · median orientation error | 0.53° -> 1.20° | Hurt | Neural Refinement for Absolute Pose Regression with Feature Synthesis (CVPR 2024 · 26 citations) |
| fuse Transformer pose with solver output | Matterport3D · mean rotation error (°) | 9.94 -> 4.93 | Helped | FAR: Flexible Accurate and Robust 6DoF Relative Camera Pose Estimation (CVPR 2024 · 17 citations) |
| add pose loss | RealEstate10K Avg · PSNR | 14.646 -> 16.03 | Helped | Unifying Correspondence Pose and NeRF for Generalized Pose-Free Novel View Synthesis (CVPR 2024 · 15 citations) |
| Generate grid-wise features from pooled features with an FC layer | COCO Keypoint val · AP | 72.5 -> 71.1 | Hurt | Spatial-Aware Regression for Keypoint Localization (CVPR 2024 · 10 citations) |
| Use CosPlace-trained features instead of ImageNet features | ShopFacade and OldHospital · median pose error | 12 / 0.45 -> 12 / 0.55; 39 / 0.73 -> 46 / 0.80 | Mixed | The Unreasonable Effectiveness of Pre-Trained Features for Camera Pose Refinement (CVPR 2024 · 10 citations) |
| Add rotation-only pre-training | 360VO-Seq4 · RRE | 0.1795 -> 0.0687 | Helped | PanoPose: Self-supervised Relative Pose Estimation for Panoramic Images (CVPR 2024 · 8 citations) |
| add depth input | CO3D, 8 images · camera center accuracy @ 0.1 | 87.8 -> 89.5 | Helped | Matrix3D: Large Photogrammetry Model All-in-One (CVPR 2025 · 6 citations) |
| Predict meta-points without support or refinement | MP-100 1-shot · PCK@0.2 | 85.31 -> 79.31 | Hurt | Meta-Point Learning and Refining for Category-Agnostic Pose Estimation (CVPR 2024 · 5 citations) |
| add the global branch | 360SPR unseen · translation error (m) | 10.33±1.3 -> 3.85±0.3 | Helped | Scene-agnostic Pose Regression for Visual Localization (CVPR 2025 · 4 citations) |
| increase evolution increments from none to five | Turning-Round · SSIM | 0.936 -> 0.947 | Helped | Direct May Not Be the Best: An Incremental Evolution View of Pose Generation (AAAI 2024 · 4 citations) |
What reviewers will expect
Datasets and baselines named in the 13 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- ARKitScenes · 4 of 13 papers
- ScanNet++ · 3 of 13 papers
- CO3D-v2 · 3 of 13 papers
- RealEstate10K · 3 of 13 papers
- ScanNet-1500 · 2 of 13 papers
- 7-Scenes · 2 of 13 papers
- MegaDepth · 2 of 13 papers
- BlendedMVS · 2 of 13 papers
Baselines
- DUSt3R · 3 of 13 papers
- PoseDiffusion · 3 of 13 papers
- MASt3R · 2 of 13 papers
- COLMAP · 2 of 13 papers
- VGGSfM · 2 of 13 papers
- RelPose · 2 of 13 papers
- RayDiffusion · 2 of 13 papers
- Marepo · 2 of 13 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 56 papers; from OpenAlex and the papers' affiliations).
- University of Oxford · 5 papers
- Zhejiang University · 3 papers
- ETH Zurich · 3 papers
- Stanford University · 3 papers
- The University of Hong Kong · 3 papers
- Carnegie Mellon University · 2 papers
- University of Toronto · 2 papers
- Czech Technical University in Prague · 2 papers
- Chinese Academy of Sciences · 2 papers
- University of Chinese Academy of Sciences · 2 papers
Use it
Download BibTeX of 40 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
- Depth-based hand pose regression (84 papers)
- Category-level object pose estimation (92 papers)
- Instance-level 6D object pose (74 papers)
- Category-level pose estimation (54 papers)
Elsewhere in the topic
- Physics and sensor-based motion capture (125 papers)
- Monocular 2D-to-3D pose lifting (90 papers)
- Transformer-based multi-view 3D pose (83 papers)
- Human shape and pose regression (80 papers)
- Temporal video pose estimation (76 papers)
- Human mesh recovery from images (58 papers)
- Multi-person pose estimation (51 papers)
- Heatmap and coordinate pose encoding (50 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).