Research map / Label-efficient and robust learning / Self-supervised and contrastive
Boosting Self-Supervised Tracking with Contextual Prompts and Noise Learning
Yaozong Zheng, Qihua Liang, Bineng Zhong, Shuimu Zeng, Yuanliang Xue, Ning Li, 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 |
|---|---|---|---|
| use semantic patch tokens instead of non-semantic queries (ablation) | PNTrack-256 · LaSOT | AUC: 64.6 -> 65.9 | Improved |
| add contextual noise (ablation) | PNTrack-256 · LaSOT | AUC: 65.1 -> 65.9 | Improved |
| add contextual prompts (ablation) | PNTrack-256 · LaSOT | AUC: 64.1 -> 65.9 | Improved |
| sample tokens with attention map instead of classification confidence map (ablation) | PNTrack-256 · LaSOT | AUC: 64.8 -> 65.9 | Improved |
| increase contextual token length from 4 to 8 (ablation) | PNTrack-256 · LaSOT | AUC: 65.4 -> 65.9 | Improved |
| increase contextual token length from 8 to 16 (ablation) | PNTrack-256 · LaSOT | AUC: 65.9 -> 65.3 | Hurt |
| start feature perturbation at epoch 75 instead of 0 (ablation) | PNTrack-256 · LaSOT | AUC: 64.5 -> 65.9 | Improved |
| start perturbation at epoch 75 instead of 50 (ablation) | PNTrack-256 · LaSOT | AUC: 65.2 -> 65.9 | Improved |
| start perturbation at epoch 100 instead of 75 (ablation) | PNTrack-256 · LaSOT | AUC: 65.9 -> 65.6 | No clear effect |
| PNTrack-256 vs SSTrack-256 on GOT10K | PNTrack-256 · GOT10K | AUC: 67.1 -> 69.0 | Improved |
| PNTrack-384 vs SSTrack-384 on TrackingNet | PNTrack-384 · TrackingNet | AUC: 80.4 -> 81.8 | Improved |
| PNTrack-384 vs SSTrack-384 on LaSOT_ext | PNTrack-384 · LaSOT_ext | AUC: 48.5 -> 49.1 | Improved |
Closest papers by meaning
- Less Is More: Token Context-Aware Learning for Object Tracking (AAAI 2025 · similarity 0.88)
- Target-Aware Tracking with Long-Term Context Attention (AAAI 2023 · similarity 0.87)
- Exploring Enhanced Contextual Information for Video-Level Object Tracking (AAAI 2025 · similarity 0.85)
- ODTrack: Online Dense Temporal Token Learning for Visual Tracking (AAAI 2024 · similarity 0.85)
- No More Shortcuts: Realizing the Potential of Temporal Self-Supervision (AAAI 2024 · similarity 0.85)
- MambaLCT: Boosting Tracking via Long-term Context State Space Model (AAAI 2025 · similarity 0.85)
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.
