The generalisation capabilities to cross and unknown scenarios are one of the main issues to solve in deepfake detection research. This work provides a new face-swap dataset generated by six algorithms, such as *Blendswap, Ghost, Simswap, Uniface, Hyperswap and Inswapper, which can be used to enhance current classifiers. Our work focuses on tackling open challenges in current datasets, such as a small number of bona fide images, a limited number of subjects, and unevenly split datasets. Our new dataset consists of around 70k high-resolution (1024×1024), subject-disjoint face images, lossy-compressed and face-swapped.
Download and Copyright:
The database will be made available to other researchers through a previously signed and licensed agreement. Interested researchers can download this database by contacting juan.tapia-farias@h-da.de and philipp.srock@h-da.de
Commercial use of the Licensed Materials is strictly prohibited without prior written permission from the Licensor. A partnership will be possible, but it requires prior coordination.
All technical reports and papers that present experimental results from this database should include an acknowledgement and a reference to [1].
(*) The face-swap algorithms must be required directly from the source.
References
[1] Philipp Srock, Juan E. Tapia, Christoph Busch: HQSwap: A Challenging High-Resolution Face-Swap Dataset for Improved Manipulation Detection, 14th International Workshop on Biometrics and Forensics Côte d’Azur, EURECOM, April 23–24, 2026. Github
Acknowledgements
This research work has been partially funded by the European Union (EU) under G.A. no. 101121280 (EINSTEIN) and CarMen (101168325), and the German Federal Ministry of Education and Research and the Hessian Ministry of Higher Education, Research, Science and the Arts within their joint support of the National Research Center for Applied Cybersecurity ATHENE.