license: cc-by-nc-sa-4.0
tags:
- anomaly-detection
- industrial-inspection
- computer-vision
- mvtec
- unsupervised-learning
MVTec Anomaly Detection Dataset (MVTec AD)
Dataset description
The MVTec Anomaly Detection (MVTec AD) dataset is a large-scale real-world dataset for unsupervised anomaly detection in industrial inspection scenarios.
It contains high-resolution images of multiple object and texture categories, including normal samples and various defect types with pixel-level ground truth masks.
Official website:
https://www.mvtec.com/company/research/datasets/mvtec-ad
Dataset format
In this repository, the dataset is provided as a single compressed archive:
mvtec_anomaly_detection.tar.xz
After downloading, the archive must be extracted locally:
tar -xJf mvtec_anomaly_detection.tar.xz
Dataset structure
Each category follows the structure:
category/
├── train/
│ └── good/
├── test/
│ ├── good/
│ └── defect_type/
└── ground_truth/
└── defect_type/
- Training set: only normal images
- Test set: normal and anomalous images
- Ground truth: pixel-level defect masks
Categories
The dataset includes objects and textures such as:
bottle, cable, capsule, carpet, grid, hazelnut, leather, metal_nut, pill, screw, tile, toothbrush, transistor, wood, zipper, and others.
License
This dataset is distributed under:
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International
(CC BY-NC-SA 4.0)
You may:
- Share and redistribute the dataset
- Adapt and build upon the dataset
Under the conditions:
- Attribution required
- Non-commercial use only
- Share-alike under the same license
Full license text:
https://creativecommons.org/licenses/by-nc-sa/4.0/
Citation
If you use this dataset in academic work, please cite:
@inproceedings{bergmann2019mvtec,
title={MVTec AD — A Comprehensive Real-World Dataset for Unsupervised Anomaly Detection},
author={Bergmann, Paul and Fauser, Michael and Sattlegger, David and Steger, Carsten},
booktitle={Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
year={2019}
}
Source and attribution
This dataset is originally provided by:
MVTec Software GmbH https://www.mvtec.com
All rights remain with the original authors.
Intended use
This dataset is intended for:
- Academic research
- Educational purposes
- Benchmarking anomaly detection algorithms
Commercial use is not permitted.
Disclaimer
This repository is not affiliated with MVTec Software GmbH. It is provided only for research and educational purposes.