Contrastive-Neuron-Pruning-for-Backdoor-Defense

This is an official implementation of the TIP 2025 Paper Contrastive-Neuron-Pruning-for-Backdoor-Defense in Pytorch.

The code is being organized and is expected to be released in March.

Citation

If you find this repo useful for your research, please consider citing our paper

@ARTICLE{10887056,
  author={Feng, Yu and Ma, Benteng and Liu, Dongnan and Zhang, Yanning and Cai, Weidong and Xia, Yong},
  journal={IEEE Transactions on Image Processing}, 
  title={Contrastive Neuron Pruning for Backdoor Defense}, 
  year={2025},
  volume={},
  number={},
  pages={1-1},
  keywords={Neurons;Training;Data models;Feature extraction;Robustness;Perturbation methods;Face recognition;Electronic mail;Computational modeling;Batch normalization;Backdoor defense;image classification;contrastive learning},
  doi={10.1109/TIP.2025.3539466}}