Research map / 3D vision / Human pose estimation
Self-Supervised Learning of Pose Embeddings From Spatiotemporal Relations in Videos
Omer Sumer, Tobias Dencker, Bjorn Ommer
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 motion-based curriculum (ablation) | temporal task, random init · Olympic Sports posture analysis | AUC: 0.592 -> 0.630 | Improved |
| add motion-based curriculum (ablation) | temporal task, ImageNet init · Olympic Sports posture analysis | AUC: 0.762 -> 0.781 | Improved |
| add motion-based curriculum (ablation) | temporal & spatial, random init · Olympic Sports posture analysis | AUC: 0.664 -> 0.679 | Improved |
| add motion-based curriculum (ablation) | temporal & spatial, ImageNet init · Olympic Sports posture analysis | AUC: 0.767 -> 0.784 | Improved |
| add spatial placement task to temporal ordering (ablation) | random init · Olympic Sports posture analysis | AUC: 0.630 -> 0.679 | Improved |
| add spatial placement task to temporal ordering (ablation) | ImageNet init · Olympic Sports posture analysis | AUC: 0.781 -> 0.784 | Improved |
| add repetition mining to temporal & spatial (ablation) | random init · Olympic Sports posture analysis | AUC: 0.679 -> 0.701 | Improved |
| add repetition mining to temporal & spatial (ablation) | ImageNet init · Olympic Sports posture analysis | AUC: 0.784 -> 0.804 | Improved |
| add repetition mining to temporal task (ablation) | random init · Olympic Sports posture analysis | AUC: 0.630 -> 0.658 | Improved |
| add repetition mining to temporal task (ablation) | ImageNet init · Olympic Sports posture analysis | AUC: 0.781 -> 0.794 | Improved |
| use spatial task alone instead of temporal alone (ablation) | random init · Olympic Sports posture analysis | AUC: 0.668 vs 0.630 | Mixed |
| self-supervised S&T fine-tuning from ImageNet AlexNet | AlexNet · Olympic Sports posture analysis | AUC: 0.722 -> 0.804 | Improved |
Closest papers by meaning
- Learning Temporal Pose Estimation from Sparsely-Labeled Videos (NeurIPS 2019 · similarity 0.88)
- Thin-Slicing for Pose: Learning to Understand Pose Without Explicit Pose Estimation (CVPR 2016 · similarity 0.88)
- 3D Human Pose Estimation Using Spatio-Temporal Networks with Explicit Occlusion Training (AAAI 2020 · similarity 0.87)
- Shuffle and Learn: Unsupervised Learning Using Temporal Order Verification (ECCV 2016 · similarity 0.87)
- SpatioTemporal Learning for Human Pose Estimation in Sparsely-Labeled Videos (AAAI 2025 · similarity 0.86)
- Emergence of a Shared Canonical Object Frame from In-the-Wild Videos (NeurIPS 2026 · 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.
