Yes. But the main issue is in the way they formulate the problem. Their output is always a transparency mask, which of course will never handle distortions.

youd have to train it to also generate and st map of the distortions but creating the ground truth version of that from the synthetic data would add a lot more to render. also its very easy to plausibly fake, its not something humans are good at seeing and knowing its wrong. you can tell its completely missing but accurate vs just distorted in a plausible way is not something most brains are tuned to notice.

Right. Things like this are why it’s difficult integrating AI into professional movie pipelines— they’re super complex in ways AI cannot (yet) replicate for very good reasons that seem superfluous or trivially replaceable by people not familiar with them.