More experiments of this workflow: Generating quick 3d models using Lumas Genie 3D. A text prompt generates 4 models in 10 seconds. Then, using Lumas webviewer, I feed a screencapture of the model into Krea ai which details the model based on another prompt. Finally, Magnific AI is used... to add the final finish. The Santa mechs are insanely refined😱 Check out the final "Renders" below. #3d #aishow more

Here's the Magnific upscaled/enhanced images.

Really enjoying seeing your workflows @MartinNebelong. Thought you may find the following interesting - using @krea_ai to augment the 3d model making process…

I am so thirsty for exportable 3d models at this quality 🤤

@MotionMark369 How long do you think it will be before we can get that from workflows like this?

Have you found using most assets fed through krea retains its visual identity? Like if i use an item graphic with specific iconography i want it to retain but to be fed through does it usually stay prominently visible?

Hey, that depends entirely on how far you push the "creativity" slider. You can stay very close to the source but if you do so most of the image will too. You could generate an image with high ai hallucinations and then apply the iconography after if you need to.

Amazing workflow - thanks again Martin 👏🏻- I guess the elephant in the room with all these new txt2model workflows will be the quality of output mesh, UV layout, topology etc.. especially the latter if you need to rig/animate/deform them predictably. 🤔. Needs another process/step - but I’m sure even this will be automated over time with AI I suspect. 😎

Love me some mecha Santas! Here’s a piece I did last year 🤣

I have an idea to make it 3d: looking at recently presented video model W.A.L.T, you can generate e 360 rotating video. Using it in Polycam or a gaussian spaltting software should be enough to extract detailed 3d models

If consistent, maybe you can make some kind of "gaussian splat density"