Research map / Detection and segmentation / Tracking
Towards Grand Unification of Object Tracking
Bin Yan, Yi Jiang, Peize Sun, Dong Wang, Zehuan Yuan, Ping Luo, Huchuan Lu
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 target prior to detection head (ablation) | Unicorn (ConvNeXt-Tiny) · LaSOT | AUC: 50.9 -> 67.7 | Improved |
| add target prior to detection head (ablation) | Unicorn (ConvNeXt-Tiny) · DAVIS-2017 | J&F: 29.2 -> 68.0 | Improved |
| add target prior to detection head (ablation) | Unicorn (ConvNeXt-Tiny) · BDD100K MOT | mMOTA: 37.6 -> 39.9 | Improved |
| fuse prior by broadcast sum instead of concatenation (ablation) | Unicorn (ConvNeXt-Tiny) · LaSOT | AUC: 66.8 -> 67.7 | Improved |
| fuse prior by broadcast sum instead of concatenation (ablation) | Unicorn (ConvNeXt-Tiny) · BDD100K MOTS | mMOTSA: 27.2 -> 29.7 | Improved |
| use deformable attention instead of convolution for feature interaction (ablation) | Unicorn (ConvNeXt-Tiny) · LaSOT | AUC: 66.8 -> 67.7 | Improved |
| use deformable attention instead of full attention (ablation) | Unicorn (ConvNeXt-Tiny) · BDD100K MOTS | mMOTSA: 26.7 -> 29.7 | Improved |
| use ConvNeXt-Tiny instead of ResNet-50 backbone (ablation) | Unicorn · LaSOT | AUC: 65.3 -> 67.7 | Improved |
| use ConvNeXt-Tiny instead of ResNet-50 backbone (ablation) | Unicorn · BDD100K MOTS | mMOTSA: 30.8 -> 29.7 | Hurt |
| unify four tasks in one model vs separate single-task models (ablation) | Unicorn (ConvNeXt-Tiny) · LaSOT | AUC: 67.5 -> 67.7 | No clear effect |
| unify four tasks in one model vs separate single-task models (ablation) | Unicorn (ConvNeXt-Tiny) · BDD100K MOTS | mMOTSA: 28.1 -> 29.7 | Improved |
| unify four tasks in one model vs separate single-task models (ablation) | Unicorn (ConvNeXt-Tiny) · DAVIS-2017 | J&F: 68.4 -> 68.0 | No clear effect |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- Siam R-CNN — Siam R-CNN: Visual Tracking by Re-Detection (CVPR 2020) · better on 1 of 2 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
- SUTrack: Towards Simple and Unified Single Object Tracking (AAAI 2025 · similarity 0.89)
- TrackingNet: A Large-Scale Dataset and Benchmark for Object Tracking in the Wild (ECCV 2018 · similarity 0.89)
- OmniTracker: Unifying Visual Object Tracking by Tracking-With-Detection (TPAMI 2025 · similarity 0.88)
- VastTrack: Vast Category Visual Object Tracking (NeurIPS 2024 · similarity 0.87)
- Towards Generalizable Multi-Object Tracking (CVPR 2024 · similarity 0.86)
- Is Multiple Object Tracking a Matter of Specialization? (NeurIPS 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.
