FaceForensics: A Large-scale Video Dataset for Forgery Detection in Human Faces

By Andreas Rössler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, Matthias Nießner. 2018

452 citations recorded by the source corpus.

With recent advances in computer vision and graphics, it is now possible to generate videos with extremely realistic synthetic faces, even in real time. Countless applications are possible, some of which raise a legitimate alarm, calling for reliable detectors of fake videos. In fact, distinguishing between original and manipulated video can be a challenge for humans and computers alike, especially when the videos are compressed or have low resolution, as it often happens on social networks. Research on the detection of face manipulations has been seriously hampered by the lack of adequate datasets. To this end, we introduce a novel face manipulation dataset of about half a million edited images (from over 1000 videos). The manipulations have been generated with a state-of-the-art face editing approach. It exceeds all existing video manipulation datasets by at least an order of magnitude. Using our new dataset, we introduce benchmarks for classical image forensic tasks, including classification and segmentation, considering videos compressed at various quality levels. In addition, we introduce a benchmark evaluation for creating indistinguishable forgeries with known ground truth; for instance with generative refinement models.

Subfields: Computer Vision and Pattern Recognition, Generative Models

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Generative Models

Citations in the corpus

Video-to-Video Synthesis (2018, methodology): “• Face video dataset [57].”
Deepfakes and Disinformation: Exploring the Impact of Synthetic Political Video on Deception, Uncertainty, and Trust in News (2020, methodology): “First, Rössler et al. (2018) show that most users have limited ability to discern between a deepfake and an authentic video, so manipulation checks would have been biased by a high degree of guessing.”
TGANv2: Efficient Training of Large Models for Video Generation with Multiple Subsampling Layers (2018, methodology): “Using the FaceForensics [39], we confirmed our model successfully generated videos with high resolution (256× 256 pixels).”
Exploiting Visual Artifacts to Expose Deepfakes and Face Manipulations (2019, methodology): “[33] proposes a large dataset containg Face2Face manipulations and the detection based on CNNs.”
Deepfake Video Detection through Optical Flow Based CNN (2019, background): “From another side, various studies have lastly focused on the problem to detect deepfake-like videos; most of them by analyzing possible inconsistencies within RGB frames of the video [9, 10, 1].”
Dynamic Neural Radiance Fields for Monocular 4D Facial Avatar Reconstruction (2020, background): “Moreover, the detection of manipulated videos is an active area of research [31, 12, 1, 24, 30, 13, 11, 2, 44] and we foresee significant improvements in this domain.”
Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Forensics (2019, methodology): “Multi-task [34] Designed CNN FaceForensics [39] https://github.”
Few-shot Video-to-Video Synthesis (2019, methodology): “We use the real videos in the FaceForensics dataset [44], which contains 854 videos of news brieﬁng from different reporters.”
Deepfake Video Detection Using Recurrent Neural Networks (2018, background): “[34], which has about half a million edited images that have been generated with feature-based face editing [38].”
FakeCatcher: Detection of Synthetic Portrait Videos using Biological Signals (2019, methodology): “The dataset is released on our project page 1 We document ω , segment count, and corresponding segment and video accuracies, on toy (FF test), entire (FF) [46], and Deep Fakes (DF) datasets.”
Media Forensics and DeepFakes: An Overview (2020, methodology): “using the approach described in [182], however this analysis can be accomplished in different ways [183].”
On the Detection of Digital Face Manipulation (2019, unknown): “[52], Deep-fake [26], and FaceForensics/FaceForensics ++ [34, 35].”
DeepFakes and Beyond: A Survey of Face Manipulation and Fake Detection (2020, methodology): “To the best of our knowledge, the only available database to date focused on facial expression manipulation is Face-Forensics++ [6], an extension of FaceForensics [61].”
Deepfakes generation and detection: state-of-the-art, open challenges, countermeasures, and way forward (2021, methodology): “This dataset was presented in 2019 as an extended form of the FaceForensics dataset [220], which contains videos with facial expressions manipulation only, and which was released in 2018.”
Capsule-forensics: Using Capsule Networks to Detect Forged Images and Videos (2018, methodology): “Moreover, while some methods can detect a single forged image [11, 12, 13], others require video input [9].”
Multi-task Learning for Detecting and Segmenting Manipulated Facial Images and Videos (2019, methodology): “The sliding window approach is used more for detecting spoofing regions generated by a computer to create spoof images or videos from bona fide ones [25, 21, 26].”
ForensicTransfer: Weakly-supervised Domain Adaptation for Forgery Detection (2018, methodology): “Original videos and Face2Face ones are of the FaceForensics dataset [34].”