Research map / 3D vision / Human pose estimation
EgoPoser: Robust Real-Time Egocentric Pose Estimation from Sparse and Intermittent Observations Everywhere
Jiaxi Jiang, Paul Streli, Manuel Meier, Christian Holz
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 |
|---|---|---|---|
| add global motion decomposition (temporal + spatial normalization) instead of mean normalization (ablation) | EgoPoser · AMASS | MPJPE (cm): 6.25 -> 4.14 | Improved |
| retain vertical position (normalize only horizontal translation to head) (ablation) | EgoPoser · AMASS | MPJPE (cm): 4.96 -> 4.45 | Improved |
| add temporal normalization on top of spatial normalization (ablation) | EgoPoser · AMASS | MPJPE (cm): 4.45 -> 4.14 | Improved |
| use head-relative local input (improved versions) instead of global input for baselines | AvatarPoser · HPS BIB_EG_Tour | MPJPE (cm): 22.53 -> 11.48 | Improved |
| use head-relative local input (improved versions) for baseline on velocity | AvatarPoser · HPS BIB_EG_Tour | MPJVE (cm/s): 60.25 -> 82.70 | Hurt |
| model FoV realistically instead of random hand masking (ablation) | EgoPoser · AMASS FoV = 90 deg | MPJPE (cm): 14.84 -> 6.60 | Improved |
| add SlowFast fusion instead of 40-frame input (ablation) | EgoPoser · AMASS | MPJVE (cm/s): 28.12 -> 25.95 | Improved |
| SlowFast vs full 80-frame input (ablation) | EgoPoser · AMASS | MPJPE (cm): 4.11 -> 4.14 | No clear effect |
| estimate body shape jointly instead of using mean shape (ablation) | EgoPoser · AMASS shape-aware | MPJPE (cm): 6.36 -> 4.79 | Improved |
| shape data augmentation plus T-pose calibration instead of mean shape (ablation) | EgoPoser · AMASS shape-aware | MPJPE (cm): 6.36 -> 5.26 | Improved |
| global position input vs offset of origin up to 50 m | AvatarPoser, AGRoL, AvatarJLM · AMASS with positional offset | position error | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- AvatarPoser — AvatarPoser: Articulated Full-Body Pose Tracking from Sparse Motion Sensing (ECCV 2022) · better on 1 of 1 shared result
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)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · 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)
- Real-Time Simulated Avatar from Head-Mounted Sensors (CVPR 2024 · similarity 0.89)
- Estimating Body and Hand Motion in an Ego-sensed World (CVPR 2025 · similarity 0.89)
- HMD-Poser: On-Device Real-time Human Motion Tracking from Scalable Sparse Observations (CVPR 2024 · similarity 0.88)
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.
