Research map / Graph learning / Graph algorithms and clustering
Learnable Graph Matching: Incorporating Graph Partitioning With Deep Feature Learning for Multiple Object Tracking
Check 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 |
|---|---|---|---|
| Replace bipartite Hungarian association with graph matching (ablation) | GMTracker · MOT17 val | IDF1: 68.1 -> 70.0 | Helped |
| Add linear interpolation (ablation) | GMTracker · MOT17 val | IDF1: 71.6 -> 73.0 | Helped |
| Add appearance feature refinement (ablation) | Graph-matching tracker · MOT17 val | IDF1: 70.2 -> 70.4 | Helped |
| Add cross-graph GCN (ablation) | Graph-matching tracker · MOT17 val | IDF1: 70.4 -> 70.6 | Helped |
| Add geometric information to cross-graph GCN (ablation) | Graph-matching tracker · MOT17 val | IDF1: 70.6 -> 71.5 | Helped |
| Train the graph-matching layer jointly with the tracker (ablation) | GMTracker · MOT17 val | IDF1: 70.2 -> 71.5 | Helped |
| Use graph matching at both training and inference instead of neither (ablation) | GMTracker · MOT17 val | IDF1: 69.5 -> 71.5 | Helped |
| Use only the last tracklet frame rather than average tracklet features (ablation) | GMTracker · MOT17 val | IDF1: 70.0 -> 69.1 | Hurt |
| Replace mean feature aggregation with moving average (ablation) | GMTracker · MOT17 val | IDF1: 70.0 vs 69.9; 70.0 vs 70.1 | No clear effect |
| Increase max age from 30 to 100 frames (ablation) | GMTracker with graph matching · MOT17 val | IDF1: 68.0 -> 70.0 | Helped |
| Increase max age from 100 to 150 frames (ablation) | GMTracker with graph matching · MOT17 val | IDF1: 70.0 -> 70.0 | No clear effect |
| Increase max age from 100 to 150 frames (ablation) | Hungarian association · MOT17 val | IDF1: 68.1 -> 67.3 | Hurt |
Closest papers by meaning
- Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking (AAAI 2024 · similarity 0.89)
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.87)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.87)
- Self-Supervised Multi-Object Tracking with Path Consistency (CVPR 2024 · similarity 0.87)
- ADA-Track: End-to-End Multi-Camera 3D Multi-Object Tracking with Alternating Detection and Association (CVPR 2024 · similarity 0.87)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · 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.