Research map / 3D vision / Human pose estimation
EgoPoseFormer v2: Accurate Egocentric Human Motion Estimation for AR/VR
Zhenyu Li, Sai Kumar Dwivedi, Filip Maric, Carlos Chacón, Nadine Bertsch, Filippo Arcadu, Tomas Hodan, Michael Ramamonjisoa, Peter Wonka, Amy Zhao, Robin Kips, Cem Keskin and 2 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| predict pose parameters with forward kinematics instead of direct keypoint head (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.35 -> 4.17 | Improved |
| add causal temporal attention (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.35 -> 4.17 | Improved |
| add projected 2D keypoint conditioning in refinement (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.30 -> 4.17 | Improved |
| condition query on auxiliary metadata (headset pose) (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.39 -> 4.17 | Improved |
| add per-joint uncertainty head (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.25 -> 4.17 | Improved |
| add auto-labeling system (pseudo-labels from 70M unlabeled frames) (ablation) | EPFv2 · EgoBody3M | wrist MPJPE (cm): 5.74 -> 5.07 | Improved |
| add auto-labeling system (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.17 -> 4.08 | Improved |
| add uncertainty distillation loss to ALS (ablation) | EPFv2 · EgoBody3M | overall MPJPE (cm): 4.08 -> 4.02 | Improved |
| EPFv2 vs EgoPoseFormer on temporal smoothness | EPFv2 · EgoBody3M | MPJVE: 0.87 -> 0.42 | Improved |
| EPFv2 w/o ALS vs Egobody3M on temporal smoothness | EPFv2 w/o ALS · EgoBody3M | MPJVE: 0.54 -> 0.42 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- EgoPoseFormer — EgoPoseFormer: A Simple Baseline for Stereo Egocentric 3D Human Pose Estimation (ECCV 2024) · better on 6 of 6 shared results
- Egobody3M — EgoBody: Human Body Shape and Motion of Interacting People from Head-Mounted Devices (ECCV 2022) · better on 6 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- FRAME: Floor-aligned Representation for Avatar Motion from Egocentric Video (CVPR 2025 · similarity 0.91)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.90)
- EgoGen: An Egocentric Synthetic Data Generator (CVPR 2024 · similarity 0.90)
- EMHI: A Multimodal Egocentric Human Motion Dataset with HMD and Body-Worn IMUs (AAAI 2025 · similarity 0.90)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · similarity 0.90)
- 3D Human Pose Perception from Egocentric Stereo Videos (CVPR 2024 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
