Research map / Detection and segmentation / Tracking
Transformer Tracking With Cyclic Shifting Window Attention
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 pixel-level attention with window attention alone (ablation) | Transformer tracker · UAV123 | AUC: 66.2 -> 54.4 | Hurt |
| add cyclic shifts to window attention (ablation) | CSWinTT · UAV123 | AUC: 54.4 -> 69.7 | Helped |
| add the spatially regularized attention mask (ablation) | CSWinTT with window attention and cyclic shifts · UAV123 | AUC: 69.7 -> 70.1 | Helped |
| add relative position encoding without the spatial mask (ablation) | CSWinTT with window attention and cyclic shifts · UAV123 | AUC: 69.7 -> 69.8 | Helped |
| combine the spatial mask and relative position encoding (ablation) | CSWinTT with window attention and cyclic shifts · UAV123 | AUC: 70.1 -> 70.5 | Helped |
| use multi-scale windows rather than one fixed window size (ablation) | CSWinTT · UAV123 | AUC: 70.0 -> 70.5 | Helped |
| remove cyclic shifts of Query (ablation) | CSWinTT computation · Tracking speed | FPS: 1.0 -> 8.2 | Helped |
| halve the shifting periods (ablation) | CSWinTT computation with Query shifts removed · Tracking speed | FPS: 8.2 -> 10.9 | Helped |
| use coordinate permutations instead of direct matrix translations for cyclic shifts (ablation) | CSWinTT computation with Query shifts removed and periods halved · Tracking speed | FPS: 10.9 -> 12.4 | Helped |
| use cyclic-shifting window attention instead of the original transformer (ablation) | CSWinTT · Tracking speed | FPS: 14.9 -> 12.4 | Hurt |
Closest papers by meaning
- Compact Transformer Tracker with Correlative Masked Modeling (AAAI 2023 · similarity 0.90)
- Target Focused Shallow Transformer Framework for Efficient Visual Tracking (AAAI 2024 · similarity 0.87)
- Cross-modulated Attention Transformer for RGBT Tracking (AAAI 2025 · similarity 0.86)
- Learning Adaptive and View-Invariant Vision Transformer for Real-Time UAV Tracking (ICML 2024 · similarity 0.85)
- Sequential Fusion Based Multi-Granularity Consistency for Space-Time Transformer Tracking (AAAI 2024 · similarity 0.85)
- UTPTrack: Towards Simple and Unified Token Pruning for Visual Tracking (CVPR 2026 · similarity 0.84)
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.