I've seen other ideas for content aware paste, similar to the harmonization Neural filter, but I'd like to see the AI-powered version. Basically, when I paste content into a document, I want to be able to "make it fit."
I'm including an example of the kind of thing that would be a real break-through:
Here we've got an office backdrop, and I've dropped a picture of a pedestal on a base into a new layer. The new object does not match the perspective and lighting of the background image. Traditionally, I'm in for a long night after a bit of perspective warping with some clone stamping and custom painting to fake bits of perspective.
What about an AI powered set of tools? Can I "paste into" a selected region and have the image scaled and redrawn with an estimation of the correct perspective and lighting? Can I merge layers, adapting one to look at home in the other? --at present, I know I cannot. I asked Generative Fill to put in a Greek Vase on a stone pedestal and I got a lampshade on a metal pole. But generative is irrelevant. I need this image, manipulated to match perspective and lighting.
I imagine someday, a generated 3D object that I can position and then "Apply" as a guide. But I'd be happy in the short term to just see some kind of low-rank adaptation features for perspective matching etc.
AdChoices