Research map / Vision-language and multimodal / Human and embodied modeling
Divert More Attention to Vision-Language Tracking
Mingzhe Guo, Zhipeng Zhang, Heng Fan, Liping Jing
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 ModaMixer to SiamCAR (ablation) | SiamCAR · LaSOT | SUC: 50.7 -> 57.6 | Improved |
| Add ASS to SiamCAR (ablation) | SiamCAR · LaSOT | SUC: 50.7 -> 52.1 | Improved |
| Add ASS to the ModaMixer-enhanced tracker (ablation) | VLT_SCAR with ModaMixer · LaSOT | SUC: 57.6 -> 65.2 | Improved |
| Replace symmetrical network with an asymmetrical network (ablation) | VLT_SCAR · LaSOT | SUC: 60.0 -> 63.9 | Improved |
| Search ModaMixer post-processing blocks instead of fixing ShuffleNet blocks (ablation) | VLT_SCAR · LaSOT | SUC: 59.1 -> 63.9 | Improved |
| Remove language description at inference (ablation) | VLT_SCAR · LaSOT | SUC: 65.2% -> 53.4% | Hurt |
| Use generated pseudo-language caption at inference (ablation) | VLT_SCAR · LaSOT | SUC: 51.6% | Hurt |
| Use template visual embedding rather than a 0-tensor for videos without language annotations (ablation) | VLT_SCAR · LaSOT_Ext | SUC: 41.2 -> 44.7 | Mixed |
| Reduce language-annotated training data from 100% to 75% (ablation) | VLT_SCAR · LaSOT | SUC: 63.9% -> 61.8% | Hurt |
| Reduce language-annotated training data from 100% to 50% (ablation) | VLT_SCAR · LaSOT | SUC: 63.9% -> 58.9% | Hurt |
| Replace BERT with GPT-2 (ablation) | VLT_SCAR · LaSOT | SUC: 63.9 -> 59.3 | Hurt |
Closest papers by meaning
- Unifying Visual and Vision-Language Tracking via Contrastive Learning (AAAI 2024 · similarity 0.90)
- ChatTracker: Enhancing Visual Tracking Performance via Chatting with Multimodal Large Language Model (NeurIPS 2024 · similarity 0.89)
- MemVLT: Vision-Language Tracking with Adaptive Memory-based Prompts (NeurIPS 2024 · similarity 0.88)
- VastTrack: Vast Category Visual Object Tracking (NeurIPS 2024 · similarity 0.88)
- OVTR: End-to-End Open-Vocabulary Multiple Object Tracking with Transformer (ICLR 2025 · similarity 0.87)
- Mono3DVLT: Monocular-Video-Based 3D Visual Language Tracking (CVPR 2025 · 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.
