Learned Equivariant Rendering without Transformation Supervision

NeurIPS 2020

Learned Equivariant Rendering without Transformation Supervision

Dec 06, 2020
|
28 views
|
Details
We propose a self-supervised framework to learn scene representations from video that are automatically delineated into objects and background. Our method relies on moving objects being equivariant with respect to their transformation across frames and the background being constant. After training, we can manipulate and render the scenes in real time to create unseen combinations of objects, transformations, and backgrounds. We show results on moving MNIST with backgrounds. Speakers: Cinjon Resnick, Krishna Jatavallabhula, Kelsey Allen, Victoria Dean

Comments
loading...