Research map / Vision-language and multimodal / Cross-modal learning
SEATrack: Simple, Efficient, and Adaptive Multimodal Tracker
Junbin Su, Ziteng Xue, Shihui Zhang, Kun Chen, Weiming Hu, Zhipeng Zhang
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 AMG-LoRA to the two-stream baseline (ablation) | SEATrack · LasHeR | PR: 51.5 -> 69.8 | Improved |
| add HMoE to the two-stream baseline (ablation) | SEATrack · LasHeR | PR: 51.5 -> 67.4 | Improved |
| add HMoE on top of AMG-LoRA (ablation) | SEATrack · DepthTrack | F-score: 57.6 -> 63.2 | Improved |
| add AMG-LoRA on top of HMoE (ablation) | SEATrack · LasHeR | PR: 67.4 -> 71.6 | Improved |
| initialize scaling factors to 1 instead of 0 (ablation) | AMG-LoRA · DepthTrack | F-score: 57.5 -> 63.2 | Improved |
| initialize scaling factors to 1 instead of 0.5 (ablation) | AMG-LoRA · DepthTrack | F-score: 58.9 -> 63.2 | Improved |
| use 2 heads per expert instead of 1 (ablation) | HMoE · DepthTrack | F-score: 58.9 -> 63.2 | Improved |
| use 2 heads per expert instead of 8 (ablation) | HMoE · LasHeR | PR: 70.6 -> 71.6 | Improved |
| increase experts to 8 attention and 16 FFN (ablation) | HMoE · DepthTrack | F-score: 63.2 -> 59.1 | Hurt |
| use HMoE instead of cross-attention fusion (ablation) | SEATrack · FPS | FPS: 41.5 -> 63.8 | Improved |
| add AMG-LoRA to cross-attention fusion (ablation) | cross-attention · LasHeR | PR: 67.6 -> 69.9 | Improved |
| use HMoE instead of MCP fusion (ablation) | SEATrack · FPS | FPS: 54.7 -> 63.8 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- SDSTrack — SDSTrack: Self-Distillation Symmetric Adapter Learning for Multi-Modal Visual Object Tracking (CVPR 2024) · better on 3 of 3 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Single-Model and Any-Modality for Video Object Tracking (CVPR 2024 · similarity 0.89)
- SDSTrack: Self-Distillation Symmetric Adapter Learning for Multi-Modal Visual Object Tracking (CVPR 2024 · similarity 0.88)
- Exploiting Multimodal Spatial-temporal Patterns for Video Object Tracking (AAAI 2025 · similarity 0.87)
- SPMTrack: Spatio-Temporal Parameter-Efficient Fine-Tuning with Mixture of Experts for Scalable Visual Tracking (CVPR 2025 · similarity 0.87)
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.87)
- Temporal Adaptive RGBT Tracking with Modality Prompt (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.
