@gunfighteronesix It is a waste subject. I’d suggest you to search for “upscale images with stable diffusion” if you wanna use the tool I use.
Improving an image with AI depends on the model you use to generate images… It can change a lot of details but you can also create great stuff from a real photo and then turn it to something different depending on you imagination !
For exemple, here I took the poster for the TV series “The last of us” (great show and games btw)
Then I generated a new image, keeping the proportions and postures of the characters but I changed the background.
Next step was reworking the faces…
… until I got something I like in terms of quality and style.
I could stop at this point. But to show you, you can also rework faces by swaping them from another picture (for exemple an actor) and integrate the face on your creation like on the picture below :
It is not easy and takes some times. I did it just to show you, but the image can still be reworked and improved.