Research map / Video / Video segmentation and tracking
TrEP: Transformer-Based Evidential Prediction for Pedestrian Intention with Uncertainty
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 TrEP instead of BiPed | PIE intention prediction · PIE | AUC: 0.90 -> 0.93 | Helped |
| use TrEP instead of MMHA | JAAD intention prediction · JAAD | accuracy: 0.84 -> 0.87 | Helped |
| use TrEP instead of PSI-Intention | PSI intention prediction · PSI | accuracy: 0.76 -> 0.83 | Helped |
| add bounding-box center coordinates (ablation) | TrEP base model · PIE, JAAD, and PSI | accuracy: PIE: 0.80 -> 0.91; JAAD: 0.79 -> 0.87; PSI: 0.72 -> 0.80 | Helped |
| add bounding-box area and length-to-width ratio (ablation) | TrEP with center coordinates · PIE | accuracy: 0.91 -> 0.89 | Hurt |
| add bounding-box area and length-to-width ratio (ablation) | TrEP with center coordinates · JAAD | accuracy: 0.87 -> 0.86 | Hurt |
| add bounding-box area and length-to-width ratio (ablation) | TrEP with center coordinates · PSI | accuracy: 0.80 -> 0.83 | Helped |
| add positional encoding (ablation) | TrEP base model · PIE | accuracy: 0.90 -> 0.91 | Helped |
| add positional encoding (ablation) | TrEP base model · JAAD | accuracy: 0.85 -> 0.87 | Helped |
| add positional encoding (ablation) | TrEP base model · PSI | accuracy: 0.81 -> 0.83 | Helped |
| reject predictions with uncertainty above 0.6 | TrEP · PIE | accuracy: 0.91 -> 0.93 | Helped |
| reject predictions with uncertainty above 0.6 | TrEP · JAAD | accuracy: 0.87 -> 0.91 | Helped |
Closest papers by meaning
- Understanding Interaction as You Need: Intention-Driven Pedestrian Behavior Prediction (AAAI 2026 · similarity 0.86)
- PSI: A Benchmark for Human Interpretation and Response in Traffic Interactions (NeurIPS 2025 · similarity 0.86)
- Social-Transmotion: Promptable Human Trajectory Prediction (ICLR 2024 · similarity 0.82)
- SocialMOIF: Multi-Order Intention Fusion for Pedestrian Trajectory Prediction (CVPR 2025 · similarity 0.82)
- BAT: Behavior-Aware Human-Like Trajectory Prediction for Autonomous Driving (AAAI 2024 · similarity 0.81)
- Towards Predicting Any Human Trajectory In Context (NeurIPS 2025 · similarity 0.81)
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.