I believe I've got character consistency worked out -
Even after all the effort I put into making my reference pictures, OpenArt.AI did, in fact, struggle with character consistency. There are two ways to generate character images - you can either do it in the character section, which has the option to create a pose with a 3D mannequin, or you can go through the regular image creator and use "omnireference" images in Flux Kontext. But, you don't get both. The regular image creator will actually produce consistent character images, if you call on the character and also upload reference images as omnireference. Perhaps if a pose image is uploaded as omnireference, it will use that for the pose as well. If it can, I'd use a real 3D program like Blender to make the pose, because OpenArt.AI's mannequin doesn't do fingers or expressions. Something to try one day.
Now it's time to work on images for Yaoi's insta.
I can't believe I'm realizing the Yaoi-Prize, but here are some tries at the Yaoi-Prize version of the holodeck -
I made a bunch of glowy ones, of which this was the best -
Looks more like SNW than TOS though. Cool, but not the right direction to go for the holodeck. It could be the computer core.
Next!
Now this is more like it! What a perfect pulp sci-fi vibe. I wish the TOS Connie had been reimagined like this, instead of an Apple product, or whatever SNW is.
Here's one that would be great for a "refit - "
Interesting, but more of a shuttle bay?
Another nice one, side bay for working on shuttles? Is that a deuterium fuel pod sticking out of the wall? Is this the Yaoi-Prize, or the JR-Prise?