DreamBooth Virtual Actors: train AI Models from 3D figures

[@cthrower]

You are amazing!!! You got a new sub for life!!!! Keep pushing the envelope!

[@NeoShameMan]

Do you plan to test image to image? that is, pose a scene with minimal elements, then use the ai + prompt to transfer the composition into a fully develop scenes?

[@CutsceneArtist]

I'm really interested in the image2image with a depth map.... I need a 2D depthmap workflow inside Unity.

[@tonywhite4476]

I applaud your initiative but this seems a little extra. First you don't need a 3D figure to train a model, you just need the images of the model. It would seem an easier workflow, since you've obviously paid hundreds of dollars for Reallusion software, if you took an image from stable diffusion and use Headshop to create a 3D character from that image. Then use the lipsync and auto-rig tools to bring it to life. I have the whole Reallusion Gambit and there's no way I'm going to use an AI image over a real 3D render. I may use it for backgrounds, art, etc. AI are good for a couple oohs and ahhs but it's nowhere near production ready...not without post-processing anyway. Btw, what did you use for lip syncing?