Draw2Image

After spending some time on GenAI image, I found my breakthrough on image composition. I’m now able to guide my composition using a drawing as input and transform my drawing into a beautiful image.

Since then, I’ve been doing some test on different subjects you can see below, but the one that’s sticked so far is the TinyArtist project where kid’s drawing are fed to my ComfyUI workflow and provide what I think is the best result for a Draw2Image approach.

Following my breakthrough on image composition, I’m now able to position multiple elements as I want. those elements though are not superb, they need more details, that would be the focus of the next door picture.

here you have 2 “mermaid” sequences, 1 sunshine version with output first and painted input and the second is the red moon version with output first too

Some other test on doors and vampires