Text to image is great, but slightly limiting. What if you added art-direction tools in order to help the AI and user sync better in the creation process?
-You could have an add lighting tool (Where you can set the desired direction, intensity, number of lights, etc.)
- A 3D "wood figure" to pose how you'd like the subjects pose to be
- A pen tool with various modes to "sketch out" roughly what the AI can fill in after (like structure, figures, horizon line, pose, expressions - something to allow the artist and AI to synergize)
- A textbox with bounding edges to further provided instruction into specific parts of the painting
Also, the initial text to image could be a starting place - it would be fantastic if you could select the direction you're liking and then launch into the detail editor as described above.
AdChoices