Research map / Vision-language and multimodal / Cross-modal learning
Local-to-global Cross-modal Coordination for Self-supervised RGB-T Tracking
Yueying Zhang, Timing Li, Bing Cao, Pengfei Zhu
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 hierarchical local coordination (HLC) (ablation) | LGCTrack baseline · LasHeR | PR: 54.8 -> 58.3 | Improved |
| add turning-phase global coordination (TGC) (ablation) | LGCTrack baseline · LasHeR | PR: 54.8 -> 55.7 | Improved |
| combine HLC and TGC (ablation) | LGCTrack · RGBT234 | PR: 78.5 -> 84.7 | Improved |
| use hierarchical collaboration instead of vanilla fusion (ablation) | LGCTrack · LasHeR | PR: 58.3 -> 58.2 (w/o gate), 59.5 (full) | Improved |
| add dynamic gating to HLC (ablation) | LGCTrack · RGBT234 | PR: 83.0 -> 84.7 | Improved |
| apply HLC across all layers 1-12 instead of a single stage (ablation) | LGCTrack · LasHeR / RGBT234 | PR: LasHeR 60.4 -> 59.5, RGBT234 83.1 -> 84.7 | Mixed |
| write refined token to the other stream instead of back to its own stream (ablation) | LGCTrack · LasHeR | PR: 58.4 -> 59.5 | Improved |
| use TGC cross-write instead of directly exchanging global tokens (ablation) | LGCTrack · LasHeR | PR: 58.9 -> 59.5 | Improved |
| update both modality global tokens instead of one (ablation) | LGCTrack · LasHeR | PR: 57.8 -> 59.5 (single, TIR or RGB: 57.8, 56.3) | Improved |
| apply TGC only at forward-to-backward transition (ablation) | LGCTrack · LasHeR | PR: 59.1 -> 59.5 | Improved |
| use LGCTrack instead of SSTrack baseline | LGCTrack · LasHeR | PR: 54.8 -> 59.5 | Improved |
| use self-supervised LGCTrack instead of fully supervised CAFormer | LGCTrack · GTOT | PR: 91.8 -> 91.9 | Improved |
Closest papers by meaning
- Cross-modulated Attention Transformer for RGBT Tracking (AAAI 2025 · similarity 0.89)
- Temporal Adaptive RGBT Tracking with Modality Prompt (AAAI 2024 · similarity 0.88)
- RGBT Tracking via All-layer Multimodal Interactions with Progressive Fusion Mamba (AAAI 2025 · similarity 0.88)
- Single-Model and Any-Modality for Video Object Tracking (CVPR 2024 · similarity 0.87)
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.87)
- SDSTrack: Self-Distillation Symmetric Adapter Learning for Multi-Modal Visual Object Tracking (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.
