search

harlanhong / CVPR2022-DaGAN

Official code for CVPR2022 paper: Depth-Aware Generative Adversarial Network for Talking Head Video Generation
https://harlanhong.github.io/publications/dagan.html
Other
958 stars 125 forks source link

Question about the background of images #2

Closed KangweiiLiu closed 2 years ago

KangweiiLiu commented 2 years ago

Thanks for this incredible work! I've looked at the demo gif on the project homepage, I was wondering about why the background is moving with head movement, is there any way to disentangle the foreground and background？

harlanhong commented 2 years ago

Hello,

Because the background of the driving video is moving with the head movement, that of the resulted video is also moving. For the disentanglement of the foreground and background, a simple way is to utilize a segmentation network to segment the human body. We're also investigating further, as we've found that when the human hair is looser, the outer hair is often treated as the background. Please continue to watch our repo, the code will be uploaded in the near future, and we will upload some new demo videos.

Best, Fa-Ting Hong