← all repositories
983632847/Awesome-Multimodal-Object-Tracking

A card catalog for the sensor soup of object tracking

Someone has to sort the flood of RGB+thermal, RGB+depth, and RGB+language tracking papers so you don't have to.

1.1k stars Jupyter Notebook Computer VisionLearning
Awesome-Multimodal-Object-Tracking
Not currently ranked — collecting fresh signals.
star history

What it does This repository is a continuously updated survey and bibliography for multi-modal single-object tracking. It catalogs papers, datasets, and benchmarks across modality pairings like RGB+thermal (RGBT), RGB+depth (RGBD), RGB+event (RGBE), and RGB+language (RGBL), plus newer niches like embodied and hyperspectral tracking. The authors also use it as a home for their own releases, including the WebUAV-3M and UW-COT220 datasets.

The interesting bit The field is a tower of Babel where every lab invents a new sensor pairing and a new acronym. The repo imposes order by grouping work into technical paradigms—self-supervised learning, prompt learning, state space models—so you can see whether a paper is actually solving a new problem or just adding a new sensor to an old pipeline. It is part reading list, part map.

Key highlights

  • Taxonomy covers RGBD, RGBT, RGBE, RGBL, RGB+sonar, RGB+NIR, and miscellaneous RGB+X tracking.
  • Surveys technical paradigms including knowledge distillation, generative models, and state space models.
  • Bundles original datasets and benchmarks: WebUAV-3M, UAV-Anti-UAV, UW-COT220, and VL-SOT500.
  • Maintains an update log with recent arXiv and conference papers through 2026.
  • Accompanies a dedicated survey paper on arXiv.

Caveats

  • This is a curated markdown list and taxonomy, not a runnable tracking framework or library.
  • Some sections, such as Embodied Visual Tracking datasets, are marked “Coming soon.”
  • Scope is strictly single-object tracking; multi-object work is out of bounds.

Verdict Grab this if you are a researcher or engineer trying to navigate the proliferation of multi-modal tracking literature and need a quick index of what datasets and papers matter. Skip it if you are hunting for a plug-and-play tracker codebase.

Frequently asked

What is 983632847/Awesome-Multimodal-Object-Tracking?
Someone has to sort the flood of RGB+thermal, RGB+depth, and RGB+language tracking papers so you don't have to.
Is Awesome-Multimodal-Object-Tracking open source?
Yes — 983632847/Awesome-Multimodal-Object-Tracking is open source, released under the MIT license.
What language is Awesome-Multimodal-Object-Tracking written in?
983632847/Awesome-Multimodal-Object-Tracking is primarily written in Jupyter Notebook.
How popular is Awesome-Multimodal-Object-Tracking?
983632847/Awesome-Multimodal-Object-Tracking has 1.1k stars on GitHub.
Where can I find Awesome-Multimodal-Object-Tracking?
983632847/Awesome-Multimodal-Object-Tracking is on GitHub at https://github.com/983632847/Awesome-Multimodal-Object-Tracking.

heatdrop uses Google Analytics to see which pages get read — nothing else. Your call. How we handle data.