Generative Radiance Fields generate 3D-consistent images, scale well to high resolution and require only unposed 2D images for training.
We analyze the trade-off between annotation time & driving policy performance for several intermediate scene representations.
A flexible implicit neural representation to perform large-scale 3D reconstruction.
We define the new task of 3D controllable image synthesis and propose an approach for solving it by reasoning both in 3D space and in the 2D image domain.
Within the first year of their life, humans develop a common-sense understanding of the physical behavior of the world. This understanding relies heavily on ...