Celeb-DF: A New Dataset for DeepFake Forensics

DeepFakes are AI-synthesized face-swapping videos which have currently become an emerging problem. A lot of AI labs and researchers are working on algorithms to detect DeepFake videos. But due to low visual quality and many artifacts, these video datasets don’t reflect the reality of DeepFake Videos circulated on the internet.

This DeepFake Forensics (Celeb-DF) dataset contains real, and DeepFake synthesized videos having similar visual quality on par with those circulated online. The Celeb-DF has 408 original videos collected from YouTube with subjects of different ages, ethnic groups, and genders, and 795 DeepFake videos synthesized from these real videos.

Paper: https://arxiv.org/pdf/1909.12962.pdf

Dataset: http://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html

Source: http://www.cs.albany.edu/~lsw/celeb-deepfakeforensics.html