Anomaly Detection via Reverse Distillation from One-Class Embedding
⭐️
⭐️
⭐️
)
Implementation (Official Code -
Environment
pytorch == 1.91
torchvision == 0.10.1
numpy == 1.20.3
scipy == 1.7.1
sklearn == 1.0
PIL == 8.3.2
-
Dataset
You should download MVTec from MVTec AD: MVTec Software. The folder "mvtec" should be unpacked into the code folder.
-
Train and Test the Model We have write both training and evaluation function in the main.py, execute the following command to see the training and evaluation results.
python main.py
Reference
@article{deng2022anomaly,
title={Anomaly Detection via Reverse Distillation from One-Class Embedding},
author={Deng, Hanqiu and Li, Xingyu},
journal={arXiv preprint arXiv:2201.10703},
year={2022}
}