Research map / Vision-language and multimodal / Cross-modal learning
Semantic Feature Purification for Adversarially-Aware RGB-T Tracking
Jiahao Wang, Fang Liu, Hao Wang, Shuo Li, Xinyi Wang, Puhua Chen
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 random perturbation injection across modalities (ablation) | SFPT · LasHeR eps=1/255 | PR: 70.1 -> 70.4 | Improved |
| add APG-CMF module (ablation) | SFPT · LasHeR eps=1/255 | PR: 66.7 -> 70.4 | Improved |
| add text information guidance (ablation) | SFPT · LasHeR eps=1/255 | PR: 66.2 -> 70.4 | Improved |
| add text-guided defense network instead of U-Net (ablation) | SFPT · LasHeR eps=1/255 | PR: 62.9 -> 70.4 | Improved |
| add text prompt generator (TPG) (ablation) | SFPT · LasHeR eps=1/255 | PR: 69.1 -> 70.4 | Improved |
| use cosine similarity loss instead of MSE (ablation) | SFPT · LasHeR eps=1/255 | PR: 69.9 -> 70.4 | Improved |
| use cosine similarity loss instead of CE (ablation) | SFPT · LasHeR eps=1/255 | PR: 69.5 -> 70.4 | Improved |
| apply SFPT defense under eps=4/255 attack on search+template | DropTrack foundation tracker · LasHeR | PR: 57.6 -> 68.4 | Improved |
| apply SFPT defense under PGD attack | DropTrack foundation tracker · LasHeR | SR: 21.3 -> 30.2 | Improved |
| apply SFPT under eps=1/255 attack on search only | DropTrack foundation tracker · LasHeR | PR: 64.0 -> 69.4 | Improved |
Closest papers by meaning
- Efficient RGB-T Tracking via Cross-Modality Distillation (CVPR 2023 · similarity 0.87)
- Local-to-global Cross-modal Coordination for Self-supervised RGB-T Tracking (ECCV 2026 · similarity 0.87)
- Cross-modulated Attention Transformer for RGBT Tracking (AAAI 2025 · similarity 0.86)
- Temporal Adaptive RGBT Tracking with Modality Prompt (AAAI 2024 · similarity 0.86)
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.86)
- Single-Model and Any-Modality for Video 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.
