Research map / Detection and segmentation / Tracking
One More Check: Making “Fake Background” Be Tracked Again
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 |
|---|---|---|---|
| Add re-check network to CSTrack (ablation) | CSTrack with boundary-aware regression · MOT17 | MOTA: 73.1 -> 76.3 | Helped |
| Add boundary-aware regression (ablation) | CSTrack · MOT17 | MOTA: 70.6 -> 73.1 | Helped |
| Use OMC instead of FairMOT | MOT16 tracking · MOT16 | MOTA: 74.9 -> 76.4 | Helped |
| Use OMC instead of FairMOT | MOT17 tracking · MOT17 | MOTA: 73.7 -> 76.3 | Helped |
| Use OMC instead of FairMOT | MOT20 tracking · MOT20 | MOTA: 61.8 -> 70.7 | Helped |
| Replace global search with local search (ablation) | Re-check network · MOT17 | MOTA: 76.3 -> 75.4 | Hurt |
| Remove response-region shrinking (ablation) | Re-check network · MOT17 | MOTA: 76.3 -> 73.6 | Hurt |
| Remove current-frame visual feature from refinement (ablation) | Re-check network · MOT17 | MOTA: 76.3 -> 69.3 | Hurt |
| Remove the inverted bottleneck module (ablation) | Re-check network · MOT17 | MOTA: 76.3 -> 75.7 | Hurt |
| Replace hard mask with Gaussian-like soft mask (ablation) | Re-check network · MOT17 | MOTA: 76.3 -> 76.0 | Hurt |
| Set response shrinking radius to 3 rather than no shrinking (ablation) | OMC · MOT15 | MOTA: 75.9 -> 77.6 | Helped |
| Lower fusion threshold to 0.1 instead of 0.5 (ablation) | OMC · MOT15 | MOTA: 77.6 -> 47.4 | Hurt |
Closest papers by meaning
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.87)
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.87)
- More Than Meets the Eye: Enhancing Multi-Object Tracking Even with Prolonged Occlusions (ICML 2025 · similarity 0.87)
- Video Diffusion Models Excel at Tracking Similar-Looking Objects Without Supervision (NeurIPS 2025 · similarity 0.86)
- Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024 · similarity 0.86)
- OneTracker: Unifying Visual Object Tracking with Foundation Models and Efficient Tuning (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.