Skip to content

CRSOT: Cross-Resolution Object Tracking using Unaligned Frame and Event Cameras

Notifications You must be signed in to change notification settings

Event-AHU/Cross_Resolution_SOT

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

26 Commits
 
 
 
 

Repository files navigation

A New Large-scale Benchmark Dataset for Cross-Resolution RGB-Event Single Object Tracking


"CRSOT: Cross-Resolution Object Tracking using Unaligned Frame and Event Cameras." Yabin Zhu, Xiao Wang, Chenglong Li, Bo Jiang, Lin Zhu, Zhixiang Huang, Yonghong Tian, Jin Tang. arXiv preprint arXiv:2401.02826 (2024). [arXiv]

💥 Update Log

  • [2024.01.08] arXiv paper and dataset are all released [arXiv] [Baidu-Disk]

🎯 Abstract

Existing datasets for RGB-DVS tracking are collected with DVS346 camera and their resolution ($346 \times 260$) is low for practical applications. Actually, only visible cameras are deployed in many practical systems, and the newly designed neuromorphic cameras may have different resolutions. The latest neuromorphic sensors can output high-definition event streams, but it is very difficult to achieve strict alignment between events and frames on both spatial and temporal views. Therefore, how to achieve accurate tracking with unaligned neuromorphic and visible sensors is a valuable but unresearched problem. In this work, we formally propose the task of object tracking using unaligned neuromorphic and visible cameras. We build the first unaligned frame-event dataset CRSOT collected with a specially built data acquisition system, which contains 1,030 high-definition RGB-Event video pairs, 304,974 video frames. In addition, we propose a novel unaligned object tracking framework that can realize robust tracking even using the loosely aligned RGB-Event data. Specifically, we extract the template and search regions of RGB and Event data and feed them into a unified ViT backbone for feature embedding. Then, we propose uncertainty perception modules to encode the RGB and Event features, respectively, then, we propose a modality uncertainty fusion module to aggregate the two modalities. These three branches are jointly optimized in the training phase. Extensive experiments demonstrate that our tracker can collaborate the dual modalities for high-performance tracking even without strictly temporal and spatial alignment.

crsot_device

🚩 Framework

framework_CRSOT.jpg

🔨 Environment

⌛ Training and Testing

📈 Experimental Results

📐 Evaluation Toolkit

📀 CRSOT Dataset Download

💾 Download from Baidu Disk:

链接:https://pan.baidu.com/s/1pUYgXQ5X-b_Yz0Hy73Io6Q?pwd=zr7j 
提取码:zr7j 

💾 Download from DropBox:

TO BE UPDATED ...

💘 Acknowledgement

[OSTrack] [CEUTrack] [VisEvent]

📰 License

MIT License

📰 Citation

If you find this work useful for your research, please cite the following papers:

@misc{zhu2024crsot,
  title={CRSOT: Cross-Resolution Object Tracking using Unaligned Frame and Event Cameras}, 
  author={Yabin Zhu and Xiao Wang and Chenglong Li and Bo Jiang and Lin Zhu and Zhixiang Huang and Yonghong Tian and Jin Tang},
  year={2024},
  eprint={2401.02826},
  archivePrefix={arXiv},
  primaryClass={cs.CV}
    }

Please leave an issue, if you have any questions about this work.

About

CRSOT: Cross-Resolution Object Tracking using Unaligned Frame and Event Cameras

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published