Research map / 3D vision / Human pose estimation
EgoXtreme: A Dataset for Robust Object Pose Estimation in Egocentric Views under Extreme Conditions
Taegyoon Yoon, Yegyu Han, Seojin Ji, Jaewoo Park, Sojeong Kim, Taein Kwon, Hyung-Sin Kim
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 |
|---|---|---|---|
| Extreme vs standard lighting (emergency scenario) | PicoPose · EgoXtreme Emergency | ADD(S) @0.3d: 67.83 -> 36.23 | Hurt |
| Add smoke (maintenance scenario) | GigaPose · EgoXtreme Maintenance | ADD(S) @0.3d: 62.77 -> 52.86 | Hurt |
| Sports scenario vs other scenarios | all pose estimators · EgoXtreme Sports | ADD(S) @0.3d: best 24.64 (GigaPose standard) | Hurt |
| Apply dehazing before pose estimation (ablation) | PicoPose · EgoXtreme Emergency | ADD(S) @0.1d: 15.12 -> 4.74 | Hurt |
| Apply dehazing before pose estimation (ablation) | PicoPose · EgoXtreme Emergency | ADD(S) @0.3d: 51.70 -> 37.74 | Hurt |
| Apply deblurring, dehazing or low-light enhancement (single) (ablation) | PicoPose · EgoXtreme Maintenance/Sports | ADD(S): maintenance @0.3d 63.32 -> about 57-58 | Hurt |
| Combine two preprocessing methods (ablation) | PicoPose · EgoXtreme Maintenance/Emergency | ADD(S) @0.3d: about -5%p (maintenance), -8%p (emergency) | Hurt |
| Direct temporal propagation instead of per-frame (ablation) | GigaPose · EgoXtreme Sports normal | ADD(S) @0.3d: 60.55 -> 14.29 (bat) | Hurt |
| Fusion temporal instead of per-frame (ablation) | GigaPose · EgoXtreme Sports normal | ADD(S) @0.3d: 17.63 -> 10.81 (pingpong); 60.55 -> 60.59 (bat) | Mixed |
| Confidence-based hybrid temporal instead of per-frame (ablation) | GigaPose · EgoXtreme Sports normal | ADD(S) @0.3d: 60.55 -> 64.46 (bat); 8.36 -> 14.35 (golf) | Improved |
| Hybrid temporal vs per-frame on pingpong (ablation) | GigaPose · EgoXtreme Sports normal (pingpong) | ADD(S) @0.3d: 17.63 -> 16.53 | No clear effect |
Closest papers by meaning
- Dynamics-regulated kinematic policy for egocentric pose estimation (NeurIPS 2021 · similarity 0.88)
- EgoSim: An Egocentric Multi-view Simulator and Real Dataset for Body-worn Cameras during Motion and Activity (NeurIPS 2024 · similarity 0.88)
- Egocentric Pose Estimation From Human Vision Span (ICCV 2021 · similarity 0.87)
- EgoPoser: Robust Real-Time Egocentric Pose Estimation from Sparse and Intermittent Observations Everywhere (ECCV 2024 · similarity 0.87)
- FRAME: Floor-aligned Representation for Avatar Motion from Egocentric Video (CVPR 2025 · similarity 0.87)
- EMHI: A Multimodal Egocentric Human Motion Dataset with HMD and Body-Worn IMUs (AAAI 2025 · 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.
