FourthM / ComAD

official code for paper entitled "Component-aware anomaly detection framework for adjustable and logical industrial visual inspection"

ComAD

This is the official code for paper entitled "Component-aware anomaly detection framework for adjustable and logical industrial visual inspection" https://arxiv.org/abs/2305.08509

Due to the randomness of KMeans, the results of each experiment will vary. For the original paper, we ran a total of five times and took the average value.

If you have any question, you could also contact [e-mail]

We will refine the code after the review.

Preparation

You need to first download MVTec Loco AD dataset https://www.mvtec.com/company/research/datasets/mvtec-loco.

Run seg_image.py

python seg_image.py --datasetpath .../mvtec_loco/

Run logical_anomaly_detection.py (make sure you have previously finished seg_image.py)

python logical_anomaly_detection.py

Acknowledgement

We use some codes from https://github.com/mhamilton723/STEGO, https://github.com/facebookresearch/dino, https://github.com/amazon-science/patchcore-inspection, and https://github.com/VitjanZ/DRAEM. A big thanks to their great work!