- cross-posted to:
- technology@lemmy.ml
You must log in or register to comment.
This brings up a good point about upscaling and other enhancements done by deep learning.
These techniques are not recovering details. They’re using their model to fabricate new details that looks plausible but do not necessarily match reality.
It’s almost as if profit driven companies and AI/ML don’t mix
It’s stuff like this that makes me smirk (being a DSLR owner).
Post processing wizardry should happen after the shot, as a manual action/enhancement not representational of the shot.