Unaligned RGBT Tracking Project

This repository contains multiple research modules related to multi-modal tracking, RGB–TIR fusion, and unaligned cross-modal UAV tracking.
Among them, our recent work:

“Progressive Multi-cue Alignment for Unaligned RGBT Tracking”
has been accepted by CVPR 2026 🎉.

“Unaligned UAV RGBT Tracking: A Largescale Benchmark and A Novel Approach”
has been accepted by AAAI 2026 🎉.

This repository includes more than this single paper, but LUART and SFCATrack are important components released here.

---

🔔 News

• 2026.02 – Our Progressive Multi-cue Alignment for Unaligned RGBT Tracking is accepted by CVPR 2026.
• 2025.11 – Our Unaligned UAV RGBT Tracking: A Largescale Benchmark and A Novel Approach is accepted by AAAI 2026.
• 2025.11 – LUART (1.02M dual-modality frames) dataset is available for download.
• Additional modules and trackers will be released soon.

---

📢 Public Release

• 2026.02, we will release the PMATrack(CVPR 2026)
• 2025.12, we publicly released:
• AAAI 2026 paper,
• SFCATrack,
• LUART dataset
• LUART evaluation toolkit

to support reproducibility and future research on unaligned RGBT tracking.

---

📦 MUART244 Dataset (Multi-platform Unaligned RGBT Tracking)

To be coming soon....

📦 LUART Dataset (Unaligned UAV RGBT Tracking)

LUART is the first large-scale benchmark focusing on unaligned UAV visible–thermal tracking.
It includes:

• 1,453 RGB–TIR sequence pairs
• 1.02M dual-modality frames
• 42 object categories
• 22 challenge attributes
• Original UAV resolutions:
  • RGB: 1920×1080
  • TIR: 640×512

📥 Download Links

LUART Dataset

• Baidu Cloud:
  https://pan.baidu.com/s/168vWYtxPqoagds8WcPuJUA
• Access Code: er4r

Evaluation Toolkit

• Baidu Cloud:
  https://pan.baidu.com/s/1lv0IBj6UtxZhj1S1UNMPsQ
• Access Code: t1vv

📦 LasHeR-Unaligned

We also provide LasHeR-Unaligned, a derived benchmark based on
LasHeR, where spatial alignment assumptions are explicitly removed to support fair evaluation of unaligned RGBT trackers.

---

📊 Benchmark Results

⭐ LUART (Test Set)

Tracker	PR ↑	NPR ↑	SR ↑
Best previous method	54.7	49.6	42.6
SFCATrack (Ours)	57.3	51.9	44.6

---

⭐ LasHeR-Unaligned

Tracker	PR ↑	NPR ↑	SR ↑
Best previous method	58.7	54.0	46.9
SFCATrack (Ours)	60.7	55.1	47.9

---

💻 Open-source Tracker

SFCATrack

Official implementation of our AAAI 2026 method:

🔗 https://github.com/Yhw-lol127/SFCATrack

---

📚 Citation

If you find this repository or the LUART dataset useful for your research,
please consider citing our AAAI 2026 paper: