Research map / Graph learning / Knowledge and temporal graphs
Spatio-Temporal Graph Transformer Networks for Pedestrian Trajectory Prediction
Cunjun Yu, Xiao Ma, Jiawei Ren, Haiyu Zhao, Shuai Yi
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 spatial Transformer (TGConv) instead of GCN (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 1.86 -> 0.41 | Improved |
| use spatial Transformer (TGConv) instead of GAT (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 0.48 -> 0.41 | Improved |
| use spatial Transformer (TGConv) instead of multi-head additive attention (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 0.56 -> 0.41 | Improved |
| use temporal Transformer instead of LSTM (no graph memory) (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 0.48 -> 0.47 | Improved |
| add graph memory (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 0.47 -> 0.41 | Improved |
| add graph memory (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | FDE: 0.97 -> 0.87 | Improved |
| use two interleaved encoders instead of a single encoder (VSTAR) (ablation) | STAR-D · ETH/UCY average (ablation, deterministic STAR) | ADE: 0.46 -> 0.41 | Improved |
| use stochastic STAR (20 samples) instead of STAR-D (ablation) | STAR · ETH/UCY average | ADE: 0.41 -> 0.26 | Improved |
| use SR-LSTM instead of simple linear regression | trajectory prediction · HOTEL | FDE: 0.72 -> 0.74 | No clear effect |
| use STAR-D instead of SR-LSTM | trajectory prediction · ETH/UCY average | ADE: 0.45 -> 0.41 | Improved |
Closest papers by meaning
- STGAT: Modeling Spatial-Temporal Interactions for Human Trajectory Prediction (ICCV 2019 · similarity 0.91)
- Social-STGCNN: A Social Spatio-Temporal Graph Convolutional Neural Network for Human Trajectory Prediction (CVPR 2020 · similarity 0.91)
- A Unified Environmental Network for Pedestrian Trajectory Prediction (AAAI 2024 · similarity 0.87)
- MS-TIP: Imputation Aware Pedestrian Trajectory Prediction (ICML 2024 · similarity 0.87)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.87)
- GigaTraj: Predicting Long-term Trajectories of Hundreds of Pedestrians in Gigapixel Complex Scenes (CVPR 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.
