Current Generative fill is a breakthrough, but can you make it even more advanced?
Here's what I mean. Now we just add a text prompt and a new object appears.
What if we need to match an image + a certain object from another image?
Not an abstract hat, glasses, watch, car, etc., but the real ones?
So, the advanced mode should make the following steps in 1 click:
match the scale, perspective, colors & lighting for a realistic composite.
Output: a layer with a mask (in case we need to make some manual adjustments).