FFHQ-Morphs

Description

This dataset contains face morphing attacks based on images from the Flickr-Faces-HQ (FFHQ) dataset as the source identities. It includes 5 different morphing algorithms (two landmark-based, two GAN-based and one diffusion-based), applied to pairs of source images consistent across every algorithm. In the related research work, we highlight the relevance of FFHQ-Morphs as training data for training morphing attack detection systems. Despite the morphing pairs being selected randomly, and the diversity of the source dataset leading to sometimes visually imperfect morphs, we observe that training on this dataset leads to better generalization to unseen source datasets. This makes the dataset potentially useful for the development of morphing attack detection targeting a plurality of passport standards.

License

FFHQ-Morphs is a dataset of human face morphs, originally created as training data for developing morphing attack detection (MAD) systems:

    Evaluating the Effectiveness of Attack-Agnostic Features for Morphing Attack Detection
    Laurent Colbois (Idiap Research Institute, UNIL), Sébastien Marcel (Idiap Research Institute, UNIL)
    https://gitlab.idiap.ch/bob/bob.paper.ijcb2024_agnostic_features_mad

FFHQ-Morphs is based on the Flickr-Faces-HQ dataset (FFHQ).
Pairs of face images have been selected from FFHQ. Several morphing algorithms have then been applied to those pairs of face images, resulting in the face morphs released in the FFHQ-Morphs dataset.

The individual images from FFHQ were originally published in Flickr by their respective authors under either Creative Commons BY 2.0, Creative Commons BY-NC 2.0, Public Domain Mark 1.0, Public Domain CC0 1.0, or U.S. Government Works license. All of these licenses allow free use, redistribution, and adaptation for non-commercial purposes. However, some of them require giving appropriate credit to the original author, as well as indicating any changes that were made to the images. The license and original author of each image are indicated in the metadata.

The FFHQ-Morphs dataset itself (including JSON metadata, download script, and documentation) is made available under Creative Commons BY-NC-SA 4.0 license by Idiap Research Institute. You can use, redistribute, and adapt it for non-commercial purposes, as long as you (a) give appropriate credit by citing our paper, (b) indicate any changes that you've made, and (c) distribute any derivative works under the same license.

Reference

If you use this dataset, please cite the following publication:

@INPROCEEDINGS{Colbois_IJCB2024_2024,
         author = {Colbois, Laurent and Marcel, S{\'{e}}bastien},
       projects = {Idiap, Biometrics Center},
          month = sep,
          title = {Evaluating the Effectiveness of Attack-Agnostic Features for Morphing Attack Detection},
      booktitle = {International Joint Conference on Biometrics},
           year = {2024},
}