Re:Experiment: Can 3D improve AI video consistency?
by numpad0 on Wednesday February 19, 2025 @09:04AM
(#43102276)
I can see how it helps to feed a 60% final product, but at that point isn't it style transfer, or, DLSS?
[ Parent
| on HN ]
Re:Experiment: Can 3D improve AI video consistency?
by fabiopolimeni on Wednesday February 19, 2025 @09:12AM
(#43102346)
Yes, the 3D data is extracted and then fed back to 2D image generation models like Flux and a relight step, at least in the examples above.