Research map / 3D vision / Human pose estimation
Overlooked Poses Actually Make Sense: Distilling Privileged Knowledge for Human Motion Prediction
Xiaoning Sun, Qiongjie Cui, Huaijiang Sun, Bin Li, Weiqing Li, Jianfeng Lu
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 |
|---|---|---|---|
| use PK-GCN instead of Traj-GCN | H3.6M · average MPJPE 1000ms | MPJPE (mm): 114.3 -> 106.4 | Improved |
| use PK-GCN instead of MSR-GCN | CMU-Mocap · average MPJPE 400ms | MPJPE (mm): 42.9 -> 40.3 | Improved |
| use PK-GCN instead of Traj-GCN | 3DPW · average MPJPE 1000ms | MPJPE (mm): 117.8 -> 114.2 | Improved |
| use PK-GCN instead of MSR-GCN | CMU-Mocap · average MPJPE 80ms | MPJPE (mm): 8.1 -> 9.4 | Hurt |
| use PK-GCN instead of MSR-GCN | H3.6M sitting, eating · MPJPE 400ms | MPJPE (mm): 40.4 -> 41.8 | Hurt |
| use PK-GCN instead of Traj-GCN | H3.6M discussion · MPJPE 400ms | MPJPE (mm): 71.7 -> 47.2 | Improved |
| add distilled privileged knowledge (vs traditional pattern TP) (ablation) | FP-Network · CMU-Mocap | MPJPE | Improved |
| add a single privileged pose (ablation) | PK-GCN · CMU-Mocap | MPJPE | Improved |
| use 5 instead of 10 privileged poses (ablation) | PK-GCN · CMU-Mocap early testpoints | MPJPE | Mixed |
| add privileged sequence loss (PSL) to baseline (ablation) | Traj-GCN · CMU-Mocap 10-25-10 (400ms) | MPJPE (mm): 52.7 -> 54.2 | Hurt |
| add privileged sequence loss (PSL) to baseline (ablation) | MSR-GCN · CMU-Mocap 10-25-10 (1000ms) | MPJPE (mm): 83.0 -> 85.6 | Hurt |
| reduce PSL weight to 0.6 (ablation) | Traj-GCN · CMU-Mocap 10-25-10 (400ms) | MPJPE (mm): 54.2 -> 53.7 | Mixed |
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.
- MSR-GCN — MSR-GCN: Multi-Scale Residual Graph Convolution Networks for Human Motion Prediction (ICCV 2021) · better on 4 of 4 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
- Progressively Generating Better Initial Guesses Towards Next Stages for High-Quality Human Motion Prediction (CVPR 2022 · similarity 0.89)
- Multi-level Motion Attention for Human Motion Prediction (IJCV 2021 · similarity 0.88)
- Investigating Pose Representations and Motion Contexts Modeling for 3D Motion Prediction (TPAMI 2022 · similarity 0.88)
- Learning Trajectory Dependencies for Human Motion Prediction (ICCV 2019 · similarity 0.88)
- ALIEN: Implicit Neural Representations for Human Motion Prediction under Arbitrary Latency (CVPR 2025 · similarity 0.87)
- Existence Is Chaos: Enhancing 3D Human Motion Prediction with Uncertainty Consideration (AAAI 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.
