FaceForensics++
An early and widely used benchmark dataset in the deepfake detection field, containing videos generated by multiple face-swap methods. Its relatively lower synthesis quality has led researchers to consider it insufficiently challenging, though it continues to serve as a standard baseline for comparison in subsequent studies.
Timeline (last 90 days)
In-the-wild benchmarks such as DFDC (Deepfake Detection Challenge) and DeeperForensics are closer to real dissemination scenarios on social media and video platforms than the earlier laboratory dataset FaceForensics++.
The forgery quality in the Celeb-DF dataset is significantly higher than that in the earlier FaceForensics++ dataset, and it was designed to address the problem of early benchmarks being too easy.
Large technology companies such as Meta and Microsoft have released deepfake detection tools and datasets for the research community, such as FaceForensics++ and the ASVspoof dataset.
AI video generation model training relies on large-scale facial action datasets such as DFDC and FaceForensics++.
All Facts (4)
In-the-wild benchmarks such as DFDC (Deepfake Detection Challenge) and DeeperForensics are closer to real dissemination scenarios on social media and video platforms than the earlier laboratory dataset FaceForensics++.
50%UnverifiedThe forgery quality in the Celeb-DF dataset is significantly higher than that in the earlier FaceForensics++ dataset, and it was designed to address the problem of early benchmarks being too easy.
50%UnverifiedLarge technology companies such as Meta and Microsoft have released deepfake detection tools and datasets for the research community, such as FaceForensics++ and the ASVspoof dataset.
50%UnverifiedAI video generation model training relies on large-scale facial action datasets such as DFDC and FaceForensics++.
50%