Possibility of inferencing on untrained people videos

michaildoukas / head2head

PyTorch implementation for Head2Head and Head2Head++. It can be used to fully transfer the head pose, facial expression and eye movements from a source video to a target identity.

MIT License

309 stars 44 forks source link

Possibility of inferencing on untrained people videos #13

Open amil-rp-work opened 3 years ago

amil-rp-work commented 3 years ago

Hey @michaildoukas @mrkoujan For extending this on new wild videos, for Head2Head++ (or Head2Head) what should one go through? So if the source and target are both new entities, then how does one go about inferencing? And it's tough to acquire more data and thus for source person only source video is there and similarly for target person

alcanunsal commented 2 years ago

@amil-rp-work any updates on this? I've been wondering if it's possible as well