Skip to content

Repository files navigation

PyTorch implementation and for TASE2024 paper, AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization.
这是图片

Download Weights of MVTec AD Dataset

Class Pre-trained Checkpoint Metric (I-AUROC,P-AUROC,I-AP,P-AP)
Bottle download (1.0, 0.988, 1.0, 0.792)
Cable download (0.996, 0.986, 0.998, 0.685)
Capsule download (0.984, 0.989, 0.997, 0.45)
Carpet download (0.998, 0.993, 0.999, 0.69)
Grid download (0.999, 0.989, 1.0, 0.378)
Hazelnut download (1.0, 0.986, 1.0, 0.567)
Leather download (1.0, 0.994, 1.0, 0.486
Metal nut download (0.995, 0.966, 0.999, 0.672)
Pill download (0.966, 0.983, 0.994, 0.697)
Screw download (0.978, 0.994, 0.993, 0.369)
Tile download (0.999, 0.962, 1.0, 0.552)
Toothbrush download (0.958, 0.989, 0.984, 0.519)
Transistor download (1.0, 0.981, 1.0, 0.771)
Wood download (0.993, 0.953, 0.998, 0.478)
Zipper download (0.986, 0.985, 0.996, 0.53)

Download Datasets

Please download MVTecAD dataset from MVTecAD dataset and BTAD dataset from BTAD dataset.

Installation

timm==0.3.2
pytoch==1.8.1

Citation

If you find this repository useful, please consider citing our work:

@article{luo2024ami,    
  title={AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization},  
  author={Luo, Wei and Yao, Haiming and Yu, Wenyong and Li, Zhengyong},  
  journal={IEEE Transactions on Automation Science and Engineering},  
  year={2024},  
  publisher={IEEE}  
}

About

[TASE 2024] AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization

Resources

Stars

32 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages