Research map / Probabilistic and scientific modeling / Molecules and proteins
TPNet: Trajectory Proposal Network for Motion Prediction
Liangji Fang, Qinhong Jiang, Jianping Shi, Bolei Zhou
Check 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 road semantic map input (ablation) | TPNet · Argoverse test | FDE: 5.29 -> 4.71 | Improved |
| Decay scores of proposals outside the drivable area (ablation) | TPNet-map · Argoverse test | DAC: 0.96 -> 0.99 | Improved |
| Generate multimodal proposals using reference lines (ablation) | TPNet-map · Argoverse test | minADE: 2.04 -> 1.61 | Improved |
| Generate multimodal proposals using reference lines (ablation) | TPNet-map · Argoverse test | minFDE: 4.23 -> 3.28 | Improved |
| Add cascade regression to end-point proposals (ablation) | TPNet · Argoverse validation | FDE: 4.01 -> 3.96 | Improved |
| Add proposal classification to the cascade-regression pipeline (ablation) | TPNet · Argoverse validation | FDE: 3.96 -> 3.88 | Improved |
| Use a 1.5 m grid interval instead of 1 m (ablation) | TPNet proposal generation, 6×6 m range · Argoverse validation | ADE: 1.75 -> 1.78 | Hurt |
| Use a 3 m grid interval instead of 1 m (ablation) | TPNet proposal generation, 6×6 m range · Argoverse validation | ADE: 1.75 -> 1.84 | Hurt |
| Expand proposal grid range from 6×6 m to 10×10 m at 245 anchors (ablation) | TPNet proposal generation · Argoverse validation | FDE: 3.87 -> 3.88 | No clear effect |
| Expand proposal grid range from 6×6 m to 20×20 m at 245 anchors (ablation) | TPNet proposal generation · Argoverse validation | FDE: 3.87 -> 3.93 | Hurt |
| Use 125 anchors rather than 245 within a 10×10 m range (ablation) | TPNet proposal generation · Argoverse validation | ADE: 1.75 -> 1.76 | No clear effect |
| Use 45 anchors rather than 245 within a 10×10 m range (ablation) | TPNet proposal generation · Argoverse validation | ADE: 1.75 -> 1.84 | Hurt |
Closest papers by meaning
- LTP: Lane-Based Trajectory Prediction for Autonomous Driving (CVPR 2022 · similarity 0.90)
- SMART: Simultaneous Multi-Agent Recurrent Trajectory Prediction (ECCV 2020 · similarity 0.88)
- HPNet: Dynamic Trajectory Forecasting with Historical Prediction Attention (CVPR 2024 · similarity 0.88)
- Large Scale Interactive Motion Forecasting for Autonomous Driving: The Waymo Open Motion Dataset (ICCV 2021 · similarity 0.88)
- HalentNet: Multimodal Trajectory Forecasting with Hallucinative Intents (ICLR 2021 · similarity 0.88)
- DeMo: Decoupling Motion Forecasting into Directional Intentions and Dynamic States (NeurIPS 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.
