Research map / Video / Video generation
Trace and Pace: Controllable Pedestrian Animation via Guided Trajectory Diffusion
Davis Rempe, Zhengyi Luo, Xue Bin Peng, Ye Yuan, Kris Kitani, Karsten Kreis, Sanja Fidler, Or Litany
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 clean-prediction guidance instead of noisy-sample guidance (ablation) | TRACE · ORCA-Maps waypoint | guidance error: 0.129 -> 0.105 | Improved |
| use gradient guidance instead of unguided sampling + filtering (ablation) | TRACE · ORCA-Maps waypoint | guidance error: 0.333 -> 0.105 | Improved |
| TRACE guidance vs VAE latent optimization | trajectory generation · ORCA-Maps waypoint | guidance error: 0.078 -> 0.105 | Hurt |
| TRACE vs VAE | trajectory generation · ORCA-Maps waypoint | realism lon acc EMD: 0.031 -> 0.013 | Improved |
| train on mixed ETH/UCY+nuScenes instead of nuScenes only (ablation) | TRACE w=-0.5 · nuScenes waypoint perturbed | guidance error: 0.977 -> 0.802 | Improved |
| classifier-free sampling w=-0.5 instead of w=0 (ablation) | TRACE mixed · nuScenes waypoint perturbed | guidance error: 1.129 -> 0.802 | Improved |
| classifier-free sampling w=-0.5 instead of w=0 (ablation) | TRACE mixed · nuScenes social groups | guidance error: 0.244 -> 0.245 | No clear effect |
| add agent-avoid guidance (ablation) | TRACE+PACER · animation flat crowd | fail rate: 0.087 -> 0.013 | Improved |
| add waypoint guidance (ablation) | TRACE+PACER · animation random terrain | fail rate: 0.093 -> 0.107 | Hurt |
| add learned value-function guidance (ablation) | TRACE waypoint-guided · animation obstacles | fail rate: 0.220 -> 0.178 | Improved |
| add learned value-function guidance (ablation) | TRACE waypoint-guided · animation flat crowd | waypoint error: 0.248 -> 0.175 | Improved |
Closest papers by meaning
- PACER+: On-Demand Pedestrian Animation Controller in Driving Scenarios (CVPR 2024 · similarity 0.88)
- Model-based Diffusion for Trajectory Optimization (NeurIPS 2024 · similarity 0.86)
- Humanoid Horizon: Extending Task Horizon in Whole-Body Loco-Manipulation via Parallel Training, Dynamic Starting, and Reward Gating (NeurIPS 2026 · similarity 0.86)
- Continuous Locomotive Crowd Behavior Generation (CVPR 2025 · similarity 0.85)
- Human Motion Diffusion as a Generative Prior (ICLR 2024 · similarity 0.85)
- H-GAP: Humanoid Control with a Generalist Planner (ICLR 2024 · similarity 0.84)
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.
