Playable Environments: Video Manipulation in Space and Time

TL;DR

This paper introduces Playable Environments, a novel framework for interactive video generation and manipulation in space and time, enabling users to control objects and viewpoints in 3D with a single image.

Contribution

It presents a new environment representation that allows real-time manipulation of objects and camera viewpoints in generated videos, using unsupervised learning and volumetric rendering techniques.

Findings

Enables interactive 3D video manipulation from a single image

Supports diverse object appearances with style-based modulation

Introduces large-scale datasets with significant camera movements

Abstract

We present Playable Environments - a new representation for interactive video generation and manipulation in space and time. With a single image at inference time, our novel framework allows the user to move objects in 3D while generating a video by providing a sequence of desired actions. The actions are learnt in an unsupervised manner. The camera can be controlled to get the desired viewpoint. Our method builds an environment state for each frame, which can be manipulated by our proposed action module and decoded back to the image space with volumetric rendering. To support diverse appearances of objects, we extend neural radiance fields with style-based modulation. Our method trains on a collection of various monocular videos requiring only the estimated camera parameters and 2D object locations. To set a challenging benchmark, we introduce two large scale video datasets with significant camera movements. As evidenced by our experiments, playable environments enable several creative applications not attainable by prior video synthesis works, including playable 3D video generation, stylization and manipulation. Further details, code and examples are available at https://willi-menapace.github.io/playable-environments-website