FSAD-Net

This repo contains the Pytorch implementation of our paper:

Few-Shot Anomaly Detection for Polyp Frames from Colonoscopy

Yu Tian, Gabriel Maicas, Leonardo Zorron Cheng Tao Pu, Rajvinder Singh, Johan W Verjans, Gustavo Carneiro.

• Accepted at MICCAI 2020.

Training

1. Pretrain a Deep Infomax encoder using only normal data.
2. Place your own data into data/ folder and start to train the FSAD network (Classifier) using the pretrained encoder as the feature extractor.

Run:

python main.py 

Citation

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

@inproceedings{tian2020few,
  title={Few-Shot Anomaly Detection for Polyp Frames from Colonoscopy},
  author={Tian, Yu and Maicas, Gabriel and Pu, Leonardo Zorron Cheng Tao and Singh, Rajvinder and Verjans, Johan W and Carneiro, Gustavo},
  booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention},
  pages={274--284},
  year={2020},
  organization={Springer}
}