Research map / Detection and segmentation / Tracking
Correlation-Embedded Transformer Tracking: A Single-Branch Framework
Fei Xie, Wankou Yang, Chunyu Wang, Lei Chu, Yue Cao, Chao Ma, Wenjun Zeng
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 |
|---|---|---|---|
| embed correlation in the feature network (single branch) instead of Siamese layer-wise aggregation with cross-attention (ablation) | Transformer tracker · GOT-10k | AO: 61.5 -> 65.0 | Improved |
| embed correlation instead of Siamese layer-wise aggregation, with DW-Corr (ablation) | Transformer tracker · GOT-10k | AO: 60.1 -> 65.9 | Improved |
| embed correlation instead of Siamese, DCF operator (ablation) | Transformer tracker · GOT-10k | AO: 30.3 -> 39.2 | Improved |
| use cross-attention instead of depth-wise correlation in Siamese transformer (ablation) | Siamese transformer tracker · GOT-10k | AO: 60.1 -> 61.5 | Improved |
| use cross-attention instead of DWC in Siamese CNN tracker (ablation) | Siamese CNN tracker · GOT-10k | AO: 56.2 -> 57.5 | Improved |
| add hierarchical structure (ablation) | SBT · GOT-10k | AO: 62.1 -> 65.9 | Improved |
| add masked image modeling pre-training (ablation) | SBT · GOT-10k | AO: 62.1 -> 71.0 | Improved |
| add temporal modeling (ablation) | SBT · GOT-10k | AO: 62.1 -> 64.2 | Improved |
| add Mix-MLP head (ablation) | SBT · GOT-10k | AO: 62.1 -> 63.5 | Improved |
| combine all four modifications (ablation) | SBT · GOT-10k | AO: 62.1 -> 74.2 | Improved |
| scale model up from 85M plain ViT to 305M (ablation) | SBT · ITB | AUC: 61.1 -> 62.4 | Improved |
| use deeper hierarchical model (23 vs 17 blocks) (ablation) | SBT · ITB | AUC: 57.5 -> 58.2 | No clear effect |
Closest papers by meaning
- Compact Transformer Tracker with Correlative Masked Modeling (AAAI 2023 · similarity 0.91)
- Correlation-Aware Deep Tracking (CVPR 2022 · similarity 0.90)
- Visual Object Tracking with Discriminative Filters and Siamese Networks: A Survey and Outlook (TPAMI 2022 · similarity 0.90)
- Unified Transformer Tracker for Object Tracking (CVPR 2022 · similarity 0.90)
- High-Performance Transformer Tracking (TPAMI 2022 · similarity 0.90)
- SMILEtrack: SiMIlarity LEarning for Occlusion-Aware Multiple Object Tracking (AAAI 2024 · similarity 0.87)
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.
