Research map / 3D vision / Human pose estimation
Instance-level 6D object pose
Predicts 6D poses of known or novel objects via iterative refinement, viewpoint encoding and few-shot matching.
74 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 Benchmarks Tried Checklist Preprints Who Export
Where it stands
Still moving The best reported result is still improving on 2 of 6 tracked benchmarks.
14 papers in 2023–24 and 22 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic grew, from 5.2% to 7.3%.
Open questions the data points to
- LM-O (AR): best went from 56.2 (up to 2024) to 80.1 in 2025 — SCFlow2: Plug-and-Play Object Pose Refiner with Shape-Constraint Scene Flow (CVPR 2025 · 0 citations)
- YCB-V (AR): best went from 69.6 (up to 2024) to 91.4 in 2025 — SCFlow2: Plug-and-Play Object Pose Refiner with Shape-Constraint Scene Flow (CVPR 2025 · 0 citations)
What looks settled
- T-LESS (AR): best up to 2024 83, best in 2025–26 65 from 3 papers, at most 0.2% better, if at all.
No longer reported by papers here: YCB-V (AUC of ADD-S, last in 2024); YCB-Video (ADD-S AUC, last in 2022); YCB-V (ADD(-S), 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 · Uncertainty-Driven 6D Pose Estimation of Objects and Scenes From a Single RGB Image (CVPR 2016 · 558 citations)
- 2017 · Global Hypothesis Generation for 6D Object Pose Estimation (CVPR 2017 · 137 citations)
- 2018 · Real-Time Seamless Single Shot 6D Object Pose Prediction (CVPR 2018 · 943 citations)
- 2019 · DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion (CVPR 2019 · 1,158 citations)
- 2020 · PVN3D: A Deep Point-Wise 3D Keypoints Voting Network for 6DoF Pose Estimation (CVPR 2020 · 611 citations)
- 2021 · GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation (CVPR 2021 · 453 citations)
- 2022 · ZebraPose: Coarse To Fine Surface Encoding for 6DoF Object Pose Estimation (CVPR 2022 · 183 citations)
- 2023 · TexPose: Neural Texture Learning for Self-Supervised 6D Object Pose Estimation (CVPR 2023 · 45 citations)
- 2024 · FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · 315 citations)
- 2025 · Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · 23 citations)
- 2026 · CoordAR: One-Reference 6D Pose Estimation of Novel Objects via Autoregressive Coordinate Map Generation (AAAI 2026 · 1 citation)
Earliest papers here
- Uncertainty-Driven 6D Pose Estimation of Objects and Scenes From a Single RGB Image (CVPR 2016 · 558 citations)
- Recovering 6D Object Pose and Predicting Next-Best-View in the Crowd (CVPR 2016 · 240 citations)
- Global Hypothesis Generation for 6D Object Pose Estimation (CVPR 2017 · 137 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.
T-LESS · AR (higher is better)
6 papers, 2020–2025 · flat since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2020 | 47.6 | EPOS: Estimating 6D Pose of Objects With Symmetries (CVPR 2020 · 247 citations) |
| 2024 | 83 | FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · 315 citations) |
YCB-V · AUC of ADD-S (higher is better)
6 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 91.6 | GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation (CVPR 2021 · 453 citations) |
| 2024 | 97.6 | HiPose: Hierarchical Binary Surface Encoding and Correspondence Pruning for RGB-D 6DoF Object Pose Estimation (CVPR 2024 · 20 citations) |
YCB-Video · ADD-S AUC (higher is better)
6 papers, 2019–2022 · not reported since 2022
| Year | Best so far | Set by |
|---|---|---|
| 2019 | 93.1 | DenseFusion: 6D Object Pose Estimation by Iterative Dense Fusion (CVPR 2019 · 1,158 citations) |
| 2022 | 95.2 | Uni6D: A Unified CNN Framework Without Projection Breakdown for 6D Pose Estimation (CVPR 2022 · 58 citations) |
LM-O · AR (higher is better)
5 papers, 2020–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2020 | 44.3 | EPOS: Estimating 6D Pose of Objects With Symmetries (CVPR 2020 · 247 citations) |
| 2024 | 56.2 | MatchU: Matching Unseen Objects for 6D Pose Estimation from RGB-D Images (CVPR 2024 · 25 citations) |
| 2025 | 80.1 | SCFlow2: Plug-and-Play Object Pose Refiner with Shape-Constraint Scene Flow (CVPR 2025 · 0 citations) |
YCB-V · AR (higher is better)
5 papers, 2020–2025 · still improving
| Year | Best so far | Set by |
|---|---|---|
| 2020 | 69.6 | EPOS: Estimating 6D Pose of Objects With Symmetries (CVPR 2020 · 247 citations) |
| 2025 | 91.4 | SCFlow2: Plug-and-Play Object Pose Refiner with Shape-Constraint Scene Flow (CVPR 2025 · 0 citations) |
YCB-V · ADD(-S) (higher is better)
5 papers, 2021–2024 · not reported since 2024
| Year | Best so far | Set by |
|---|---|---|
| 2021 | 60.1 | GDR-Net: Geometry-Guided Direct Regression Network for Monocular 6D Object Pose Estimation (CVPR 2021 · 453 citations) |
| 2022 | 80.5 | ZebraPose: Coarse To Fine Surface Encoding for 6DoF Object Pose Estimation (CVPR 2022 · 183 citations) |
| 2024 | 83.8 | 6D-Diff: A Keypoint Diffusion Framework for 6D Object Pose Estimation (CVPR 2024 · 39 citations) |
What has been tried
62 papers here report 619 measured changes (a component added, removed or swapped, compared with a baseline): helped 78%, no clear effect 3%, hurt 16%, 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 LLM texture augmentation | YCB-Video · AUC of ADD | 90.83 -> 91.52 | Helped | FoundationPose: Unified 6D Pose Estimation and Tracking of Novel Objects (CVPR 2024 · 315 citations) |
| add the diffusion denoising process rather than predict keypoints directly from heatmaps | LM-O · ADD(-S) | 49.2 -> 79.6 | Helped | 6D-Diff: A Keypoint Diffusion Framework for 6D Object Pose Estimation (CVPR 2024 · 39 citations) |
| use ground-truth NOCS maps with sensor depth instead of predicted NOCS maps | HouseCat6D test · 3D IoU (25% / 50%) | 50.0 / 21.2 -> 96.7 / 93.6 | Helped | HouseCat6D - A Large-Scale Multi-Modal Category Level 6D Object Perception Dataset with Household Objects in Realistic Scenarios (CVPR 2024 · 34 citations) |
| Use GMM-based pose hypothesis sampling instead of naïve sampling | BOP datasets (LM-O, T-LESS, TUD-L, IC-BIN, YCB-V) · Average Recall | 22.5 -> 31.3; 22.5 -> 35.8 | Helped | GenFlow: Generalizable Recurrent Flow for 6D Pose Refinement of Novel Objects (CVPR 2024 · 30 citations) |
| Add BCM Loss | BOP-5 · mean AR | 48.1 -> 50.5 | Helped | MatchU: Matching Unseen Objects for 6D Pose Estimation from RGB-D Images (CVPR 2024 · 25 citations) |
| Remove all object alignment steps | HO3D · ADD | 40.4 -> 0.0 | Hurt | Any6D: Model-free 6D Pose Estimation of Novel Objects (CVPR 2025 · 23 citations) |
| Use hierarchical correspondence pruning instead of direct Kabsch pose solving | LM-O · ADD(-S) recall (%) | 86.65 -> 89.62 | Helped | HiPose: Hierarchical Binary Surface Encoding and Correspondence Pruning for RGB-D 6DoF Object Pose Estimation (CVPR 2024 · 20 citations) |
| Use soft rather than hard pose-class labels | T-LESS · AR | 74.8 -> 77.1; increase of 2.3% | Helped | MRC-Net: 6-DoF Pose Estimation with MultiScale Residual Correlation (CVPR 2024 · 16 citations) |
| Use MASt3R rather than DINOv2 features | BOP datasets · AR | Pos3R 32.3, 47.3, 53.9, 31.5, 33.1, 25.1, 53.7; DINOv2 (L11) 18.5, 25.0, 22.5, 14.2, 15.6, 14.1, 32.1 | Helped | Pos3R: 6D Pose Estimation for Unseen Objects Made Easy (CVPR 2025 · 12 citations) |
| Use joint SE(3) diffusion instead of separate R3SO(3) diffusion | T-LESS · R@2 | 27.98 -> 47.21 | Helped | Confronting Ambiguity in 6D Object Pose Estimation via Score-Based Diffusion on SE(3) (CVPR 2024 · 11 citations) |
| replace ROC prediction with rotation and translation heads | YCB-Video ablation · ADD | 78.6 -> 59.4 | Hurt | One2Any: One-Reference 6D Pose Estimation for Any Object (CVPR 2025 · 9 citations) |
| use Mertens HDR instead of no HDR | IPD, spotlight · Median MVD | 3.271 -> 3.061 | Helped | Towards Co-Evaluation of Cameras HDR and Algorithms for Industrial-Grade 6DoF Pose Estimation (CVPR 2024 · 9 citations) |
What reviewers will expect
Datasets and baselines named in the 14 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- HO3D · 3 of 14 papers
- Toyota-Light · 3 of 14 papers
- REAL275 · 3 of 14 papers
- T-LESS · 3 of 14 papers
- YCB-Video · 3 of 14 papers
- LM-O · 2 of 14 papers
- MegaPose synthetic dataset · 2 of 14 papers
- BOP LM-O · 2 of 14 papers
Baselines
- GigaPose · 7 of 14 papers
- MegaPose · 6 of 14 papers
- FoundPose · 5 of 14 papers
- Oryon · 4 of 14 papers
- GenFlow · 4 of 14 papers
- ZS6D · 3 of 14 papers
- FoundationPose · 3 of 14 papers
- OSOP · 3 of 14 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 36 papers; from OpenAlex and the papers' affiliations).
- NA VER LABS · 2 papers
- ETH Zurich · 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
- Depth-based hand pose regression (84 papers)
- Learned camera pose estimation (96 papers)
- Category-level object pose estimation (92 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).