Research map / 3D vision / Human pose estimation
Estimating Egocentric 3D Human Pose in the Wild With External Weak Supervision
Jian Wang, Lingjie Liu, Weipeng Xu, Kripasindhu Sarkar, Diogo Luvizon, Christian Theobalt
Check 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 |
|---|---|---|---|
| combine egocentric and external-view constraints in spatio-temporal pseudo-label optimization | 3D pose pseudo-label generation · Wang et al. dataset | PA-MPJPE: 57.19 vs 68.13 | Improved |
| train on EgoPW pseudo-labels and use the proposed external-view feature supervision and domain adaptation | single-frame egocentric pose estimation · Wang et al.’s test dataset | PA-MPJPE: 81.71 vs 89.67 | Improved |
| train on EgoPW pseudo-labels and use the proposed external-view feature supervision and domain adaptation | single-frame egocentric pose estimation · Mo2Cap2 test dataset | PA-MPJPE: 83.17 vs 86.85 | Improved |
| add external-view supervision for pseudo-label generation (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | PA-MPJPE: 90.05 -> 81.71 | Improved |
| add adversarial external-feature representation learning (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | PA-MPJPE: 85.46 -> 81.71 | Improved |
| add adversarial domain adaptation between synthetic and real images (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | PA-MPJPE: 84.22 -> 81.71 | Improved |
| use the full method rather than unsupervised domain adaptation alone (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | PA-MPJPE: 91.56 -> 81.71 | Improved |
| add external-view supervision for pseudo-label generation (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | BA-MPJPE: 68.99 -> 64.87 | Improved |
| add adversarial external-feature representation learning (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | BA-MPJPE: 67.01 -> 64.87 | Improved |
| add adversarial domain adaptation between synthetic and real images (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | BA-MPJPE: 66.48 -> 64.87 | Improved |
| use the full method rather than unsupervised domain adaptation alone (ablation) | egocentric pose estimation network · Wang et al.’s test dataset | BA-MPJPE: 69.17 -> 64.87 | Improved |
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.
- VIBE — VIBE: Video Inference for Human Body Pose and Shape Estimation (CVPR 2020) · better on 2 of 2 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
- Scene-Aware Egocentric 3D Human Pose Estimation (CVPR 2023 · similarity 0.92)
- Estimating Egocentric 3D Human Pose in Global Space (ICCV 2021 · similarity 0.91)
- SelfPose: 3D Egocentric Pose Estimation From a Headset Mounted Camera (TPAMI 2020 · similarity 0.90)
- 3D Human Pose Perception from Egocentric Stereo Videos (CVPR 2024 · similarity 0.88)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.87)
- Estimating Ego-Body Pose from Doubly Sparse Egocentric Video Data (NeurIPS 2024 · similarity 0.86)
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.
