I'm having a problem with overlapping faces. about simswap-inference-pytorch HOT 3 OPEN

I made some changes in simswap.py and got better results:
soft_face_mask, _ = self.smooth_mask(face_mask)
img_white=soft_face_mask*255.
# Only take face area from the swapped image
swapped_img = swapped_img * soft_face_mask

However, in some cases, overlaps can occur.

from simswap-inference-pytorch.

Hey @mbeytekin , you can blend faces separately. At the end of the simswap.py file do not sum img_mask and target_image, instead blend each crop independently.

I think a better solution would be computing intersection between crops and compensating it during the blend step. I'll take a look later.

from simswap-inference-pytorch.

Fixed the issue with the latest update

before

After

from simswap-inference-pytorch.