Research map / Label-efficient and robust learning / Recognition and long-tail
Whose Track Is It Anyway? Improving Robustness to Tracking Errors With Affinity-Based Trajectory Prediction
Xinshuo Weng, Boris Ivanovic, Kris Kitani, Marco Pavone
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 |
|---|---|---|---|
| Use detections and affinity matrices instead of tracklets as prediction inputs (ablation) | Affinipred · nuScenes IDS | minADES: 1.792 -> 1.106 | Improved |
| Use detections and affinity matrices instead of tracklets as prediction inputs (ablation) | Affinipred · nuScenes global | minADES: 1.143 -> 0.977 | Improved |
| Add GT matching during training (ablation) | Affinipred · nuScenes IDS | minADES: 2.464 -> 1.792 | Improved |
| Add tracking-error data augmentation (ablation) | Affinipred · nuScenes IDS | minADES: 2.705 -> 2.464 | Improved |
| Add data imputation (ablation) | Affinipred · nuScenes IDS | minADES: 3.408 -> 2.705 | Improved |
| Use affinity-based prediction instead of tracklet-based prediction | trajectory prediction · nuScenes global | minADES: 2.320 -> 0.977 | Improved |
| Use affinity-based prediction instead of the standard tracklet pipeline | trajectory prediction · nuScenes IDS | minFDES: 13.892 -> 1.584 | Improved |
| Use affinity-based prediction instead of MTP | trajectory prediction · KITTI IDS | minADES: 0.707 -> 0.516 | Improved |
| Use affinity-based prediction instead of MTP | trajectory prediction · nuScenes FRAG | minADES: 7.697 -> 4.486 | Improved |
Closest papers by meaning
- Towards Trajectory Forecasting From Detection (TPAMI 2023 · similarity 0.90)
- Scene Transformer: A unified architecture for predicting future trajectories of multiple agents (ICLR 2022 · similarity 0.87)
- M2I: From Factored Marginal Trajectory Prediction to Interactive Prediction (CVPR 2022 · similarity 0.87)
- Prioritizing Perception-Guided Self-Supervision: A New Paradigm for Causal Modeling in End-to-End Autonomous Driving (NeurIPS 2025 · similarity 0.86)
- Improving Transferability for Cross-Domain Trajectory Prediction via Neural Stochastic Differential Equation (AAAI 2024 · similarity 0.86)
- CaDeT: a Causal Disentanglement Approach for Robust Trajectory Prediction in Autonomous Driving (CVPR 2024 · similarity 0.85)
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.
