By the way just blurring faces from demonstrations or actions might not be enough.
Given a low-resolution input image, model generates high-resolution images that are perceptually realistic and downscale correctly.
😺GitHub: https://github.com/tg-bomze/Face-Depixelizer📙Colab: https://colab.research.google.com/github/tg-bomze/Face-Depixelizer/blob/master/Face_Depixelizer_Eng.ipynb
P.S. Colab is based on thehttps://github.com/adamian98/pulse
@missqarnstein This is not reconstructing the original face, just showing one possible face that could get downsized to the low-rez picture. If you run the software twice with the same input you'll get two different results. Nothing to fear here.
This is a private instance. Sorry.
Go check this list of instances or this website to find one suitable to your needs. :)