Heads from behind and zooming in and out

Want to understand the training process better? Got tips for which model to use and when? This is the place for you

Forum rules

Read the FAQs and search the forum before posting a new topic.

This forum is for discussing tips and understanding the process involved with Training a Faceswap model.

If you have found a bug are having issues with the Training process not working, then you should post in the Training Support forum.

Please mark any answers that fixed your problems so others can find the solutions.

User avatar
Posts: 8
Joined: Thu May 27, 2021 3:53 am
Answers: 1
Been thanked: 1 time

Heads from behind and zooming in and out

Post by aolvera27 »

Does it benefit the model to have "faces" extracted from behind the face? I'm talking about pictures where you would be able to see the edges of the chin and the nose protruding next to the cheek. Is this information usefull, or won't it be considered when swapping faces?

Also, I get that pictures where the whole face is visible are good because there is information about the whole structure, but pictures with a little zoom are good too because then the extracted face will have more detail. At least I remember reading about it. If so, would it be a good idea to have the same picture twice, one with a little zoom and one without?

User avatar
Posts: 1495
Joined: Fri Jul 12, 2019 12:53 am
Answers: 127
Has thanked: 51 times
Been thanked: 287 times

Re: Heads from behind and zooming in and out

Post by torzdf »

To the former, it really depends on your target swap. Profiles tend to be tricky at the best of times, extreme profiles are often not necessary,

As to the latter. No, passing in the same face zoomed in and out would be unnecessary.

My word is final