MNW Benchmark Dataset
Microsoft-Northwestern-WITNESS (MNW) benchmark for evaluating AI-generated media and deepfake detection across images, video, and audio.
Notice
This dataset is intended for evaluation purposes only. It cannot be used for training or commercial purposes.
Contents
| Folder | Description |
|---|---|
AI_Images/ |
AI-generated images from many generators (~15k+ samples) |
AI_Video/ |
AI-generated video samples |
Deepfake_Audio/ |
Deepfake / AI audio samples |
Deepfake_Video/ |
Deepfake video samples |
AI_media_in_the_wild/ |
Real-world suspicious / AI-manipulated media |
docs/ |
Documentation |
mnw_ai_images_metadata.parquet |
Image metadata (path, label, generator name) |
Metadata
The parquet file contains 15,250 image entries with columns:
path— relative/local file pathlabel— typicallyfakedataset_name— generator or subset name (e.g.Adobe_Firefly_v2)
Citation
@ARTICLE{11479406,
author={Roca, Thomas and Postiglione, Marco and Gao, Chongyang and Gortner, Isabel and Wojciak, Zuzanna and Wang, Pengce and Alimardani, Mahsa and Anlen, Shirin and White, Kevin and Ferres, Juan Lavista and Kraus, Sarit and Gregory, Sam and Subrahmanian, V. S. and Murugesan, San},
journal={IEEE Intelligent Systems},
title={{The Microsoft-Northwestern-WITNESS Benchmark for Deepfake Detection}},
year={2026},
volume={41},
number={02},
pages={15-23},
doi={10.1109/MIS.2026.3668398}
}
Links
- Downloads last month
- 3