Masks and gradients

I’m starting to love art again. When I was in middle grade I knew I would grow up to be a cartoonist. Then in high school I decided writing was my calling. But I still drew things up until my fourth decade. Now I draw very little.

I had the same problem I was suffering in my writing: Blank page syndrome. I just couldn’t see anymore what I wanted to draw.

These AI-generated images just might be a back door into drawing again. I’m starting with a completely painted canvas. Then I start making simple changes. Next thing you know, I might be using AI just to suggest ideas and create assets.

For example, today I was thinking about combining several images that were slightly different, but generated from the same seed. In one image the girl has three legs. In another the lighthouse has a weird pattern. But overall, the composition is the same. I can layer each of these and use transparency masks to layer specific parts from each image.

Then I wondered if I could generate random transparency masks with Stable Diffusion. The answer is yes. I can also generate random color gradients. So I had the AI create a grayscale mask of cut up pieces of paper which I used to combine two images of a girl walking on a beach as a proof of concept.

Then I layered gradients and applied masks to make a geometric pattern that looks like it could be the back to a playing card or cover to a journal. I’m a step closer to creating the assets I need for the story-related card deck I want to create.