Closed HoiM closed 2 years ago
Well, it's simple:
Well, it's simple:
1. According to the pairs determined by FaceForensics++, the face on video-A frame-N is swapped to video-B frame-N 2. If this process fails, we use another (most) frontal face in video-A as the source, and swap it to video-B
Thank you for your reply!
Hi, thank you for your great work.
I have some questions regarding your FaceForensics++ experiments.
Can you let me know how you define the source and target pairs after getting the 10000 aligned faces?