AutoRF: Learning 3D Object Radiance Fields from Single View Observations...
Conversation
Replying to
Don't see how without predefined models. Assumptions of symmetry? Still, very cool concept... is how we fill in the gaps
1
Replying to
Right? I always look to how human minds do this. If you look at an object you've never seen before from one perspective, and then try to render the full shape in your mind.
We also have a collected set of intutions about spaces and the shape of things.
1
1
Show replies