Research map / Detection and segmentation / Tracking
NOOUGAT Towards Unified Online and Offline Multi-Object Tracking
Benjamin Missaoui, Orcun Cetintas, Guillem Brasó, Tim Meinhardt, Laura Leal-Taixé
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 |
|---|---|---|---|
| replace Hungarian matching with the learned ALT layer (ablation) | NOOUGAT online · DanceTrack val | IDF1: 65.8 -> 69.0 | Improved |
| replace heuristic stitching of SUSHI with learned ALT layer (ablation) | offline tracker · DanceTrack val | IDF1: 72.0 -> 76.6 | Improved |
| add motion cues to geometry cues (ablation) | NOOUGAT online · DanceTrack val | IDF1: 61.9 -> 64.4 | Improved |
| add appearance cues to geometry and motion (ablation) | NOOUGAT online · DanceTrack val | IDF1: 64.4 -> 69.0 | Improved |
| use appearance cues only instead of geometry only (ablation) | NOOUGAT online · DanceTrack val | IDF1: 61.9 -> 58.1 | Hurt |
| add VDC edge feature (ablation) | NOOUGAT offline · DanceTrack val | IDF1: 73.5 -> 74.9 | Improved |
| add skip gap in training (ablation) | NOOUGAT offline · DanceTrack val | IDF1: 74.9 -> 75.2 | Improved |
| add past-tracks augmentation (ablation) | NOOUGAT offline · DanceTrack val | IDF1: 75.2 -> 76.6 | Improved |
| increase processing stride from online to offline (T=256) (ablation) | NOOUGAT · DanceTrack test | AssA: 54.9 -> 58.7 | Improved |
| increase processing stride from online to offline (T=256) (ablation) | NOOUGAT · SportsMOT test | AssA: 74.3 -> 83.0 | Improved |
| increase retention window beyond 30 frames (ablation) | NOOUGAT online · SportsMOT val | performance | Improved |
| share weights between ALT layer and hierarchy (ablation) | NOOUGAT · training | iterations to converge: 90k -> 40k | Improved |
Closest papers by meaning
- Focusing on Tracks for Online Multi-Object Tracking (CVPR 2025 · similarity 0.88)
- Samba: Synchronized Set-of-Sequences Modeling for Multiple Object Tracking (ICLR 2025 · similarity 0.86)
- Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking (AAAI 2024 · similarity 0.86)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · similarity 0.86)
- Hybrid-SORT: Weak Cues Matter for Online Multi-Object Tracking (AAAI 2024 · similarity 0.86)
- Sampling-Resilient Multi-Object Tracking (AAAI 2024 · similarity 0.86)
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.
