Research map / Label-efficient and robust learning / Recognition and long-tail
Multi-Task Driven Feature Models for Thermal Infrared Tracking
Qiao Liu, Xin Li, Zhenyu He, Nana Fan, Di Yuan, Wei Liu, Yongsheng Liang
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 the auxiliary classification branch (ablation) | MMNet · VOT-TIR2015 | EAO: 0.282 -> 0.307 | Improved |
| Add the auxiliary classification branch (ablation) | MMNet · VOT-TIR2017 | EAO: 0.254 -> 0.274 | Improved |
| Add the holistic correlation module to fine-grained matching (ablation) | MMNet with Dis + Cls · VOT-TIR2015 | Acc: 0.51 -> 0.58 | Improved |
| Add the holistic correlation module to fine-grained matching (ablation) | MMNet with Dis + Cls · VOT-TIR2017 | Acc: 0.52 -> 0.55 | Improved |
| Add the pixel-level correlation module to fine-grained matching (ablation) | MMNet with Dis + Cls · VOT-TIR2015 | Acc: 0.51 -> 0.57 | Improved |
| Add the pixel-level correlation module to fine-grained matching (ablation) | MMNet with Dis + Cls · VOT-TIR2017 | Acc: 0.52 -> 0.56 | Improved |
| Use both holistic and pixel-level correlation modules (ablation) | MMNet · VOT-TIR2015 and VOT-TIR2017 | EAO: 0.322 -> 0.332; 0.296 -> 0.320 | Improved |
| Mix VID2015 grayscale and TIR training data (ablation) | MMNet · VOT-TIR2015 | EAO: 0.332 -> 0.344 | Improved |
| Mix VID2015 grayscale and TIR training data (ablation) | MMNet · PTB-TIR | Suc: 0.502 -> 0.539; 0.519 -> 0.539 | Improved |
| Retrain the overall network on TIR after VID pretraining (ablation) | MMNet · VOT-TIR2015 and PTB-TIR | EAO / Suc: 0.332 -> 0.300; 0.502 -> 0.521 | No clear effect |
| Fine-tune VID initialization on TIR while freezing early layers and fine-grained branch (ablation) | MMNet · PTB-TIR | Suc: 0.502 -> 0.525 | Improved |
| Use MMNet rather than CFNet | TIR tracking · PTB-TIR | Suc: +10% relative gain | Improved |
Closest papers by meaning
- Collaborating Vision, Depth, and Thermal Signals for Multi-Modal Tracking: Dataset and Algorithm (NeurIPS 2025 · similarity 0.86)
- RT-RMOT: A Dataset and Framework for RGB-Thermal Referring Multi-Object Tracking (ECCV 2026 · similarity 0.86)
- Robust Collaborative Discriminative Learning for RGB-Infrared Tracking (AAAI 2018 · similarity 0.86)
- Single-Model and Any-Modality for Video Object Tracking (CVPR 2024 · similarity 0.83)
- CSTrack: Enhancing RGB-X Tracking via Compact Spatiotemporal Features (ICML 2025 · similarity 0.82)
- Pseudo Visible Feature Fine-Grained Fusion for Thermal Object Detection (CVPR 2025 · similarity 0.82)
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.
