Seconds before opening reddit i had the idea to try to create impasto depth maps from photos / diffusion model outputs.

What I came to in the few seconds of given thought was pretty close to the additive approach as far as I can tell.

Unfortunately I don't know the first thing about painting so I'll never implement that.