ENHANCE! ‘Photo-Realistic’ Emojis and Emotes With Progressive Face Super-Resolution by iforcedabot (I Forced a Bot...)
Progressive Face Super-Resolution via Attention to Facial Landmark arxiv.org is a machine learning model trained to reconstruct face images from tiny 16×16 pixel input images, scaling them up …

I guess it’s time to talk about pixelated images any why that may be no longer enough: https://iforcedabot.com/photo-realistic-emojis-and-emotes-with-progressive-face-super-resolution/

Astonishing what this machine can compute from a 16×16 pile of pixels.