Research map / Detection and segmentation / Tracking
An Efficient Token Compression Framework for Visual Object Tracking
Weijing Wu, Qihua Liang, Bineng Zhong, Haiying Xia, Zhiyi Mo, Shuxiang Song
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 ATC module to baseline without HIBlock (ablation) | ETCTrack-B224 · LaSOT | AUC: 73.7 -> 74.4 | Improved |
| add HIBlock to baseline without ATC (ablation) | ETCTrack-B224 · LaSOT | AUC: 73.7 -> 74.4 | Improved |
| add ATC module on top of HIBlock (ablation) | ETCTrack-B224 · LaSOT | AUC: 74.4 -> 74.9 | Improved |
| add ATC and HIBlock together (ablation) | ETCTrack-B224 · TNL2K | AUC: 59.9 -> 61.3 | Improved |
| compress template tokens with keep ratio 0.9 (ablation) | ETCTrack-B224 · LaSOT | AUC: 74.4 -> 74.9 | Improved |
| lower keep ratio to 0.5 (ablation) | ETCTrack-B224 · LaSOT | AUC: 74.4 -> 74.3 | No clear effect |
| lower keep ratio from 1.0 to 0.5 (ablation) | ETCTrack-B224 · MACs | MACs (G): 35.9 -> 28.2 | Improved |
| increase template frames from 3 to 6 (ablation) | ETCTrack-B224 · LaSOT | AUC: 73.9 -> 75.2 | Improved |
| add more template frames without ATC beyond 5 (ablation) | baseline without ATC · LaSOT | AUC | Hurt |
| use ETCTrack on ViT-B backbone (ablation) | ViT-B tracker · LaSOT | AUC: 69.9 -> 73.4 | Improved |
| use ETCTrack on Fast-iTPN-B backbone (ablation) | Fast-iTPN-B tracker · LaSOT | AUC: 72.8 -> 74.9 | Improved |
| use first 8 layers of Fast-iTPN as TCM instead of ViT-B from scratch (ablation) | ETCTrack-B224 · LaSOT | AUC: 74.2 -> 74.9 | Improved |
Closest papers by meaning
- ODTrack: Online Dense Temporal Token Learning for Visual Tracking (AAAI 2024 · similarity 0.89)
- UTPTrack: Towards Simple and Unified Token Pruning for Visual Tracking (CVPR 2026 · similarity 0.89)
- Learning Channel-Aware Metric Attention for Harmonised Online Visual Object Tracking (TPAMI 2026 · similarity 0.88)
- Less Is More: Token Context-Aware Learning for Object Tracking (AAAI 2025 · similarity 0.88)
- MIMTrack: In-Context Tracking via Masked Image Modeling (AAAI 2025 · similarity 0.87)
- OneTracker: Unifying Visual Object Tracking with Foundation Models and Efficient Tuning (CVPR 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.
