Awesome
AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization (IEEE TASE 2024)
PyTorch implementation and for TASE2024 paper, AMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization.
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
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}
}