Equivariant Neural Rendering

less than 1 minute read

Published:

learned representation transforms like a real 3D scene

  • require relative transformations
  • train in a style transfer way