OpenForensics Dataset

OpenForensics dataset is poses a high level of challenges in localizing forged faces among multiple human faces in in-the-wild scenes. This dataset is designed with face-wise rich annotations explicitly for face forgery detection and segmentation. With its rich annotations, OpenForensics dataset has great potentials for research in both deepfake prevention and general human face detection.

Version 1.0.0 consists of 115K images with 334K human faces. This dataset contains various backgrounds and multiple people of various ages, genders, poses, positions, and face occlusions.

Presentation (ICCV 2021)

Downloads

If you use our datasets, codes or results, please cite the following publications:

Publications

  • Trung-Nghia Le, Huy H. Nguyen, Junichi Yamagishi, Isao Echizen, "OpenForensics: Large-Scale Challenging Dataset For Multi-Face Forgery Detection And Segmentation In-The-Wild", ICCV, 2021. [PDF] [Presentation] [Poster]

License

Our dataset is published under the CC-BY 4.0 License.