Research map / Detection and segmentation / Tracking
MotionTrack: Learning Robust Short-Term and Long-Term Motions for Multi-Object Tracking
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 the Interaction Module to the baseline tracker (ablation) | MotionTrack baseline · MOT17 validation | IDF1: 82.6 -> 83.0 | Helped |
| Add the Interaction Module to the baseline tracker (ablation) | MotionTrack baseline · MOT17 validation | AssA: 72.4 -> 72.9 | Helped |
| Add the Refind Module after adding the Interaction Module (ablation) | MotionTrack · MOT17 validation | IDF1: 83.0 -> 83.7 | Helped |
| Add the Refind Module after adding the Interaction Module (ablation) | MotionTrack · MOT17 validation | HOTA: 70.5 -> 70.8 | Helped |
| Use history-trajectory-based refinding and error compensation instead of iterative-prediction IoU refinding (ablation) | MotionTrack · MOT17 validation, occlusion limit 30 to 120 frames | IDF1: 82.6 -> 83.3 | Helped |
| Raise the allowed occlusion duration from 30 to 120 frames (ablation) | IoU-based tracker · MOT17 validation | IDF1: 80.9 -> 80.1 | Hurt |
| Raise the allowed occlusion duration from 30 to 120 frames (ablation) | ReID-based tracker · MOT17 validation | IDF1: 77.2 -> 70.4 | Hurt |
| Use MotionTrack rather than the baseline for crowded or occluded cases | MOT17 validation · Visibility < 0.25, minimum duration ≥ 20 | MOTA: 77.2 -> 78.3 | Helped |
| Use MotionTrack rather than the baseline for crowded or occluded cases | MOT17 validation · Visibility < 0.25, minimum duration ≥ 100 | MOTA: 71.5 -> 72.4 | Helped |
| Use MotionTrack rather than P3AFormer(+W&B) | MOT17 test · MOT17 test | MOTA: 81.2 -> 81.1 | Hurt |
| Use MotionTrack rather than P3AFormer(+W&B) | MOT20 test · MOT20 test | IDF1: 76.4 -> 76.5 | Helped |
Closest papers by meaning
- DiffusionTrack: Diffusion Model for Multi-Object Tracking (AAAI 2024 · similarity 0.90)
- UCMCTrack: Multi-Object Tracking with Uniform Camera Motion Compensation (AAAI 2024 · similarity 0.89)
- Sampling-Resilient Multi-Object Tracking (AAAI 2024 · similarity 0.89)
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.89)
- Multiple Object Tracking as ID Prediction (CVPR 2025 · similarity 0.88)
- DiffMOT: A Real-time Diffusion-based Multiple Object Tracker with Non-linear Prediction (CVPR 2024 · similarity 0.88)
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.