From their research abstract:
The resulting 3D model of the given text can be viewed from any angle, relit by arbitrary illumination, or composited into any 3D environment. Our approach requires no 3D training data and no modifications to the image diffusion model, demonstrating the effectiveness of pretrained image diffusion models as priors.
Well, that didn’t take long to escalate. I foresee some issues for the architecture and design professions in the near future.
1 Like
seems like a premature judgement to me. Having a look at all the sites providing “free and paid 3d models” for 5 dollars, looking at their tags “printable” etc and the handability of AI generators, I honestly doubt that just from shadows all the important aspects of a complex project like a house or mowing jewelry can be taken in account. The models in the video look more like phone application made scans of general objects. It’s still very impressive though. Maybe it will be usable as a tool, but won’t replace us completely in the next few years.
1 Like
thanks thanks thanks, I can’t wait to have a look at it when I get home!
1 Like