/FSBA

Few-Shot Backdoor Attacks on Visual Object Tracking

This is the official implementation of our paper Few-Shot Backdoor Attacks on Visual Object Tracking, accepted by the International Conference on Learning Representations (ICLR), 2022. This research project is developed based on Python 3 and Pytorch, created by Yiming Li and Haoxiang Zhong.

Pipeline

Pipeline

Codes

Please download our submitted supplementary codes (for reproducing the tracking process) here. The codes for reproducing main experiments including the training process will be released soon.

Reference

If our work or this repo is useful for your research, please cite our paper as follows:

@inproceedings{li2022few,
  title={Few-Shot Backdoor Attacks on Visual Object Tracking},
  author={Li, Yiming and Zhong, Haoxiang and Ma, Xingjun and Jiang, Yong and Xia, Shu-Tao},
  booktitle={ICLR},
  year={2022}
}