Research map / Detection and segmentation / Tracking
Spatial-Temporal Relation Networks for Multi-Object Tracking
Jiarui Xu, Yue Cao, Zheng Zhang, Han Hu
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 |
|---|---|---|---|
| add cosine similarity to unary appearance feature (ablation) | appearance feature · 2DMOT2015 (train/val split, ablation) | MOTA: 19.8 -> 29.8 | Improved |
| use cosine value alone instead of unary appearance (ablation) | appearance feature · 2DMOT2015 (train/val split, ablation) | MOTA: 19.8 -> 25.2 | Improved |
| add unary location features (ablation) | A · 2DMOT2015 (train/val split, ablation) | MOTA: 29.8 -> 31.7 | Improved |
| add motion location features (ablation) | A · 2DMOT2015 (train/val split, ablation) | MOTA: 29.8 -> 31.0 | Improved |
| add both unary and motion location features (ablation) | A · 2DMOT2015 (train/val split, ablation) | MOTA: 29.8 -> 32.3 | Improved |
| add both location features (ablation) | A · 2DMOT2015 (train/val split, ablation) | ID switches: 515 -> 129 | Improved |
| add spatial relation reasoning (ablation) | A+L · 2DMOT2015 (train/val split, ablation) | MOTA: 32.3 -> 34.8 | Improved |
| add spatial relation reasoning (ablation) | A+L · 2DMOT2015 (train/val split, ablation) | FP: 2,004 -> 947 | Improved |
| add spatial relation reasoning (ablation) | A+L · 2DMOT2015 (train/val split, ablation) | IDF1: 47.1 -> 46.5 | Hurt |
| add temporal relation reasoning (ablation) | A+L+S · 2DMOT2015 (train/val split, ablation) | MOTA: 34.8 -> 36.2 | Improved |
| add temporal relation reasoning instead of average over frames (ablation) | A+L+S · 2DMOT2015 (train/val split, ablation) | MOTA: 33.1 -> 36.2 | Improved |
| add temporal relation reasoning instead of max-pooling over frames (ablation) | A+L+S · 2DMOT2015 (train/val split, ablation) | MOTA: 33.9 -> 36.2 | Improved |
Closest papers by meaning
- Multiple Object Tracking With Correlation Learning (CVPR 2021 · similarity 0.89)
- Robust Object Tracking Based on Temporal and Spatial Deep Networks (ICCV 2017 · similarity 0.89)
- MOTR: End-to-End Multiple-Object Tracking with TRansformer (ECCV 2022 · similarity 0.88)
- GRAE-3DMOT: Geometry Relation-Aware Encoder for Online 3D Multi-Object Tracking (CVPR 2025 · similarity 0.88)
- Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking (AAAI 2024 · similarity 0.87)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.87)
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.
