Research map / 3D vision / Human pose estimation
Long-term Human Motion Prediction with Scene Context
Zhe Cao, Hang Gao, Karttikeya Mangalam, Qi-Zhi Cai, Minh Vo, Jitendra Malik
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 |
|---|---|---|---|
| represent 3D path as depth plus 2D heatmap instead of direct xyz regression (ablation) | PathNet · GTA-IM | mean MPJPE (mm): 191 mm vs. 173 mm | Improved |
| add scene image input (ablation) | Ours deterministic · GTA-IM | mean MPJPE (mm): 196 -> 173 | Improved |
| use full image instead of masked human crop (ablation) | Ours deterministic · GTA-IM | mean MPJPE (mm): 189 -> 173 | Improved |
| add depth image (RGBD) to input (ablation) | Ours deterministic · GTA-IM | mean MPJPE (mm): 173 -> 172 | No clear effect |
| use ground-truth destination instead of predicted | Ours deterministic · GTA-IM | mean MPJPE (mm): 173 -> 137 | Improved |
| sample 10 destinations with GoalNet (best of samples) | Ours stochastic · GTA-IM | mean MPJPE (mm): 173 -> 165 | Improved |
| sample 4 destinations with GoalNet | Ours stochastic · GTA-IM | mean MPJPE (mm): 173 -> 173 | No clear effect |
| sample 3 destinations with GoalNet | Ours stochastic · PROX | mean MPJPE (mm): 270 mm vs. 264 mm | Improved |
| sample 10 destinations with GoalNet | Ours stochastic · PROX | mean MPJPE (mm): 270 -> 249 | Improved |
| add pretraining on GTA-IM (ablation) | Ours deterministic · PROX | mean MPJPE (mm): 280 -> 270 | Improved |
| use ground-truth destination instead of predicted | Ours deterministic · PROX | mean MPJPE (mm): 270 -> 237 | Improved |
| sample 30 destinations (min error) for 2D goal prediction | GoalNet · GTA-IM 2D goal | 2D MPJPE (px): 23.7 -> 12.2 | Improved |
Closest papers by meaning
- Context-Aware Human Motion Prediction (CVPR 2020 · similarity 0.91)
- Scene-aware Human Motion Forecasting via Mutual Distance Prediction (ECCV 2024 · similarity 0.91)
- FutureHuman3D: Forecasting Complex Long-Term 3D Human Behavior from Video Observations (CVPR 2024 · similarity 0.89)
- Harmonizing Stochasticity and Determinism: Scene-responsive Diverse Human Motion Prediction (NeurIPS 2024 · similarity 0.89)
- MotionMap: Representing Multimodality in Human Pose Forecasting (CVPR 2025 · similarity 0.88)
- SimpliHuMoN: Simplifying Human Motion Prediction (NeurIPS 2026 · 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.
