QUEEN: QUantized Efficient ENcoding of Dynamic Gaussians for Streaming Free-viewpoint Videos

The dynamic world that we perceive around us is not 2D, but rather 3D. Unlike 2D videos, which are ubiquitous, the question of how to effectively capture, encode and disseminate free-viewpoint videos (FVV) of dynamic 3D scenes, which can be viewed at any instance of time and from any viewpoint, has intrigued computer vision and graphics researchers for much time.