Commenting on Tuesday’s “AI and Me” post, Lily wrote, “Fascinating! You must start with some reality or is it like drawing a portrait? May we see STANA alongside her avatar please?”
I don’t know what you would call my process, but I simply start with a bare-bones command, for example, “70-year-old balding male wearing glasses, Red Sox T-shirt, jeans, sneakers, in Fenway Park.”
After AI generates an image according to my initial command, I will examine the image, determine what changes are necessary and send a new command based on those changes.
The first or second command usually is not exactly what I want and I may repeat tweaking the image three, four, five or even more times. (Yesterday’s image of me in the Red Sox T-shirt was a rare one-command success!)
Lily’s wish is my command and the image below is me and my avatar side-by-side. I picked a photo where I was wearing an outfit and hairdo similar to my AI image. However, I had none wearing eyeglasses, so I generated a pair via AI and pasted the eyeglasses onto my photograph.
![]() |
Wearing Boston Proper |
![]() |
Marius Goring femulating in French television’s The O Agency Files (1968). |
No comments:
Post a Comment