Out of couple thousand extractions from a variety of videos, there are a few dozen or so images where the target subject's face is blurry (out of focus or zoomed in or motion blurred) but faceswap obviously identified the face and added it to the alignments file. Should I delete these and "clean" the alignments file? Or does having some blurry images actually help the model?
Should I remove blurry faces from extracts?
Read the FAQs and search the forum before posting a new topic.
This forum is for discussing tips and understanding the process involved for Extracting and preparing face sets for training a model in Faceswap.
If you have found a bug are having issues with the Extraction process not working, then you should post in the Extract Support forum.
Please mark any answers that fixed your problems so others can find the solutions.
Re: Should I remove blurry faces from extracts?
I find that leaving blurry faces in helps with initial training, but I try to remove most of them for the final "no warp" training to bring out more detail (though I'm not sure if this actually helps much)