zlacker

[return to "RealFill: Image completion using diffusion models"]
1. pedalp+iX[view] [source] 2023-09-30 01:13:15
>>flavor+(OP)
Mix this with some nerf/guasian spat, or other 3D rendering and we have photos where you can re-frame after being shot. No more selfie sticks, perhaps use both cameras at one time to capture more of a scene for infill.

Some will say "but that isn't a real photo of what was there", but our memories of what was in a photo or a scene aren't perfect anyway.

◧◩
2. drumtt+201[view] [source] 2023-09-30 01:49:47
>>pedalp+iX
I like the idea of a 3d generated scene we can explore with VR
[go to top]