IGD

This repo contains the Pytorch implementation of our paper:

Deep One-Class Classification via Interpolated Gaussian Descriptor

Yuanhong Chen*, Yu Tian*, Guansong Pang, Gustavo Carneiro.

• Accepted at AAAI 2022 (Oral).

Dataset

Please download the MVTec AD dataset

Please download the Hyper-Kvasir Anomaly Detection Dataset from this link.

Train and Test IGD

After the setup, simply run the following command to train/test the global/local model:

./job.sh

Citation

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

@inproceedings{chen2022deep,
  title={Deep one-class classification via interpolated gaussian descriptor},
  author={Chen, Yuanhong and Tian, Yu and Pang, Guansong and Carneiro, Gustavo},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  volume={36},
  number={1},
  pages={383--392},
  year={2022}
}