Research map / Detection and segmentation / Tracking
Improving Multiple Pedestrian Tracking by Track Management and Occlusion Handling
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 |
|---|---|---|---|
| Add occlusion-aware association of detections to occluded tracks (ablation) | TMOH · MOT17 train | IDF1: +1.1 points | Helped |
| Replace occlusion handling with appearance-based re-identification (ablation) | TMOH · MOT17 train | IDF1: 1.3 points lower | Hurt |
| Add regression of inactive tracks (ablation) | TMOH · MOT17 train | MOTA: +1.5 points | Helped |
| Add regression of inactive tracks (ablation) | TMOH · MOT17 train | IDF1: +0.8 points | Helped |
| Delete tracks moving toward the image boundary when they leave the scene (ablation) | TMOH · MOT17 train | MOTA | Helped |
| Add the motion model and camera motion compensation (ablation) | TMOH · MOT17 train | IDF1: 72.4 -> 76.6 | Helped |
| Increase occlusion distance tolerance α above 1.2 (ablation) | TMOH occlusion handling · MOT17 train | IDF1: 0.2 points lower when α is doubled from 1 | Hurt |
| Vary occlusion distance tolerance α across evaluated values (ablation) | TMOH occlusion handling · MOT17 train | MOTA: deviation of 0.1 points | No clear effect |
| Estimate velocity from five previous bounding boxes (ablation) | TMOH motion model · MOT17 train | IDF1 | Helped |
| Track in both temporal directions and merge tracklets (ablation) | TMOH · MOT17 train | MOTA: 75.6 -> 80.8 | Helped |
| Track in both temporal directions and merge tracklets (ablation) | TMOH · MOT17 train | IDF1: 76.8 -> 78.5 | Helped |
| Track backwards instead of forwards (ablation) | TMOH · MOT17 train | MOTA: 73.5 -> 75.6 | Helped |
Closest papers by meaning
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.86)
- Probabilistic Tracklet Scoring and Inpainting for Multiple Object Tracking (CVPR 2021 · similarity 0.86)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.86)
- TGFormer: Transformer with Track Query Group for Multi-Object Tracking (AAAI 2025 · similarity 0.85)
- DeconfuseTrack: Dealing with Confusion for Multi-Object Tracking (CVPR 2024 · similarity 0.84)
- Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024 · similarity 0.83)
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.