AI image generation

Hello,
I am using Babylon.js for pose and 3D layers as part of biggest AI image editor:

It will be available very soon as well (free version + commercial Photoshop plugin).

11 Likes

Sick :fire:
Is that using controllnet?

2 Likes

Weird. Why do I get the feeling that this looks like a painting from David Hockney? :thinking:


1 Like

It is a combination of ControlNet and T2I Adapter

Absolutely - but with woman instead of men :wink:

@imperator this is impressive !!! can not wait to try it !!!

cc @PirateJC

1 Like

@imperator this is seriously cool!

@RaananW needs to see this and will probably want to talk with you. :slight_smile:

1 Like

Crazy cool!! congrats!!1

1 Like

:slightly_smiling_face:write me on my Discord to try out now: sd-photoshop-oss (I am “imperator” there)

Thanks for the feedback! :slightly_smiling_face:
I have more examples but not uploaded to YouTube yet.
One thing I want to add is length Gizmo in combination with that rotation gizmo which does not work at the moment and I stuck with workaround…
So that would be the skeleton only but could used for animals like giraffe where bone length are different.

1 Like

@Cedric should be able to work back on it as soon as the release is done :slight_smile:

LoL. It’s only now I saw your prompt actually asking for a style based on David’s work. I swear I didn’t see it before. Which likely makes me a knowledgeable person in the history of art and also… a poor reader :grin: :face_with_hand_over_mouth:
Well, I gotta say, since this is the prompt, the result in terms of style and use of the color palette is quite impressive.
This shall be a big ‘YES’ from me :+1: :heart_eyes:

1 Like

haha yes it is very small font size - he is one of my favourite artists btw. :slight_smile:

1 Like

Beautiful! what an awesome way to use babylon and ControlNet!
Are you doing any depth manipulations (like ZoeDepth-style), or is the babylon support mainly focused on pose generations?

1 Like

Thank you. there are two editor here: also one for 3D objects and it is shown in video (the rabbit). In this case the depth map is directly used from Babylon. There is also a detection of depth possible from image using ZoeDepth - in thios case Babylon is not used.

1 Like

release video is out showing some more examples:

3 Likes