I wrote this Generative Fill prompt to create the image in the oval:
"A dense rose garden with a stone wall and a white gate, an arbor with climbing roses over the gate, and a view of trees and sky behind the wall"
The shapes are very simple and rough. Illustrator does provide a 'Detail' slider, but even at maximum detail, it does not create a recognizable illustration of a rose garden, although it is charming. Also, there is no arbor, so my prompt failed and needs to be re-worded.
My Generative Pattern prompt for the background: "small green leaves in 3 shades on a darker green background"
The Pattern Fill is very successful. It is very easy to create a simple, pleasing pattern for use in backgrounds or packaging, and they are seamless!
Four examples of 32 images generated for the prompt "An orange cat is sitting on a windowsill, looking out the window at rooftops." Most of the cats were not generated looking out the window. AI defaults to forward-facing subjects. Even after I emphasized "back view" AI still got confused about which way the ears and whiskers should face (top row and bottom right)
My prompt: A group of graduating college students in purple robes and mortarboards with gold tassels, with Manhattan in the background. This image does not appear to be based on a photographic reference. The people are very crude, like cut-out paper dolls, and are completely unconvincing, even as a simple illustration.
Copyright © 2018 Li Gardiner - All Rights Reserved.