Research map / Detection and segmentation / Tracking
GLoMOT: Efficient Online GNN-based Low-Frame-Rate Multi-Object Tracker
Yaxuan Hu, Jie Hua, Gang Wu, Yuhong Yang, Atsushi Suzuki, Zhongyuan Wang
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 |
|---|---|---|---|
| add the full model (NM, SMP, PD) over none of them (ablation) | GLoMOT · DanceTrack val (n=10) | HOTA: 38.91 -> 44.22 | Improved |
| add full model over none of the modules (ablation) | GLoMOT · MOT17 val (n=10) | HOTA: 73.47 -> 75.02 | Improved |
| add position-only pseudo-depth (SparseTrack style) (ablation) | GLoMOT · DanceTrack val (n=2) | IDF1: 52.45 -> 51.63 | Hurt |
| add fused pseudo-depth (size+pos+occ) instead of none (ablation) | GLoMOT · DanceTrack val (n=2) | IDF1: 52.45 -> 53.33 | Improved |
| add fused pseudo-depth (ablation) | GLoMOT · DanceTrack val (n=10) | HOTA: 43.63 -> 43.68 | No clear effect |
| add fused pseudo-depth (ablation) | GLoMOT · MOT17 val (n=10) | HOTA: 74.56 -> 74.75 | Improved |
| add size-only pseudo-depth (ablation) | GLoMOT · DanceTrack val (n=2) | HOTA: 53.56 -> 54.14 | Improved |
| train on DanceTrack instead of MOT17 and test on MOT17 | GLoMOT · MOT17 (cross-domain) | HOTA: 75.91 -> 76.01 | Improved |
| online GNN with node buffer pool versus offline/other trackers | GLoMOT · MOT17 val (HOTA drop n=2 to n=10) | Delta H: 3.1 vs 6.4 (DiffMOT) | Improved |
| compare GLoMOT to MOTIP at larger frame gap | GLoMOT · DanceTrack val (n=10) | HOTA: 44.2 vs 48.6 (MOTIP) | Hurt |
| GLoMOT on high-frame-rate MOT17 test versus DiffMOT | GLoMOT · MOT17 test | HOTA: 63.3 vs 64.5 (DiffMOT) | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- DiffMOT — DiffMOT: A Real-time Diffusion-based Multiple Object Tracker with Non-linear Prediction (CVPR 2024) · better on 5 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- NOOUGAT Towards Unified Online and Offline Multi-Object Tracking (IJCV 2026 · similarity 0.88)
- Unifying Short and Long-Term Tracking With Graph Hierarchies (CVPR 2023 · similarity 0.88)
- Multi-Scene Generalized Trajectory Global Graph Solver with Composite Nodes for Multiple Object Tracking (AAAI 2024 · similarity 0.88)
- GNN3DMOT: Graph Neural Network for 3D Multi-Object Tracking With 2D-3D Multi-Feature Learning (CVPR 2020 · similarity 0.87)
- Learnable Graph Matching: Incorporating Graph Partitioning With Deep Feature Learning for Multiple Object Tracking (CVPR 2021 · similarity 0.87)
- Track-On: Transformer-based Online Point Tracking with Memory (ICLR 2025 · 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.
