The RealTimeFaceSwap-10k video call deepfake detection dataset is a dataset used to detect deepfake videos in video conferencing scenarios. This dataset includes various application scenarios and data types, aiming to provide fundamental data support for video spoofing detection.

Dataset composition:

• It contains 1,636 targeted video clips, all of which are YouTube videos licensed under Creative Commons.
• 2,000 source photos used for face swapping.
• 9,772 deepfake videos generated using face-swapping models.

Data Fields:

• Video file information includes video ID, start and end times, filename, size, and tags.
• Photo file information includes filename, tags, and source link.
• The generated deepfake video information includes the target video file name, the source photo file name, the generator model name, and additional characteristics.