Trigger without Trace: Towards Stealthy Backdoor Attack on Text-to-Image Diffusion Models

This repository contains artifacts and code related to the paper: [Trigger without Trace: Towards Stealthy Backdoor Attack on Text-to-Image Diffusion Models]

Code: https://github.com/Robin-WZQ/TwT

This study introduces TwT, an attack method based on syntactic structures that exhibits strong resistance to advanced detection methods.

πŸ“„ Citation

If you find this project useful in your research, please consider cite:

@misc{zhang2025twt,
      title={Trigger without Trace: Towards Stealthy Backdoor Attack on Text-to-Image Diffusion Models}, 
      author={Jie Zhang and Zhongqi Wang and Shiguang Shan and Xilin Chen},
      year={2025},
      eprint={2503.17724},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2503.17724}, 
}
Downloads last month
17
Safetensors
Model size
0.1B params
Tensor type
F32
Β·
Inference Providers NEW
This model isn't deployed by any Inference Provider. πŸ™‹ Ask for provider support

Paper for RobinWZQ/backdoor_KMMD_len_20_a_motor