Research map / Detection and segmentation / Tracking
Towards More Flexible and Accurate Object Tracking With Natural Language: Algorithms and Benchmark
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 AdaSwitcher to the tracker instead of using the baseline tracker alone (ablation) | SiamRPN++ (AlexNet) · TNL2K | precision: 0.344 -> 0.355 | Helped |
| Add AdaSwitcher to the tracker instead of using the baseline tracker alone (ablation) | SiamRPN++ (AlexNet) · TNL2K | success: 0.353 -> 0.370 | Helped |
| Use AdaSwitcher instead of response-score-only switching (ablation) | SiamRPN++ with TANet · TNL2K | precision: 0.347 -> 0.355 | Helped |
| Use AdaSwitcher instead of response-score-only switching (ablation) | SiamRPN++ with TANet · TNL2K | success: 0.362 -> 0.370 | Helped |
| Add frame attention to AdaSwitcher (ablation) | SiamRPN++ with AdaSwitcher · TNL2K | precision: 0.353 -> 0.355 | Helped |
| Add frame attention to AdaSwitcher (ablation) | SiamRPN++ with AdaSwitcher · TNL2K | success: 0.369 -> 0.370 | Helped |
| Add spatial coordinates to visual grounding (ablation) | Visual grounding module · TNL2K | precision: 0.103 -> 0.143 | Helped |
| Add spatial coordinates to visual grounding (ablation) | Visual grounding module · TNL2K | success: 0.124 -> 0.159 | Helped |
| Add predicted BBox history to AdaSwitcher inputs (ablation) | AdaSwitcher · TNL2K | precision: 0.350 -> 0.355 | Helped |
| Add result-image history to AdaSwitcher inputs (ablation) | AdaSwitcher · TNL2K | precision: 0.345 -> 0.355 | Helped |
| Set the switch threshold to 0.7 rather than 0.5 (ablation) | AdaSwitcher · TNL2K | precision: 0.350 -> 0.355 | Helped |
| Use language and BBox with TANet global search rather than GTI | Joint language-and-BBox tracking · LaSOT | precision: 0.47 -> 0.55 | Helped |
Closest papers by meaning
- ChatTracker: Enhancing Visual Tracking Performance via Chatting with Multimodal Large Language Model (NeurIPS 2024 · similarity 0.90)
- Unifying Visual and Vision-Language Tracking via Contrastive Learning (AAAI 2024 · similarity 0.89)
- MVLM: Template-Free Tracking via Vision-Language Margin Confidence and Memory-Gated Tracking (CVPR 2026 · similarity 0.88)
- SUTrack: Towards Simple and Unified Single Object Tracking (AAAI 2025 · similarity 0.87)
- VastTrack: Vast Category Visual Object Tracking (NeurIPS 2024 · similarity 0.87)
- OVT-B: A New Large-Scale Benchmark for Open-Vocabulary Multi-Object Tracking (NeurIPS 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.