Layered Neural Rendering for Retiming People in Video

TL;DR

This paper introduces a learning-based layered video representation that enables retiming, editing, and removing people in videos with realistic quality, by decomposing frames into separate layers for each individual.

Contribution

It presents a novel layered neural rendering method that disentangles and retimes multiple people in videos, including scene effects, for realistic editing.

Findings

Achieves high-quality retiming of complex actions

Automatically disentangles scene effects like shadows and reflections

Supports realistic removal and editing of individuals in videos

Abstract

We present a method for retiming people in an ordinary, natural video -- manipulating and editing the time in which different motions of individuals in the video occur. We can temporally align different motions, change the speed of certain actions (speeding up/slowing down, or entirely "freezing" people), or "erase" selected people from the video altogether. We achieve these effects computationally via a dedicated learning-based layered video representation, where each frame in the video is decomposed into separate RGBA layers, representing the appearance of different people in the video. A key property of our model is that it not only disentangles the direct motions of each person in the input video, but also correlates each person automatically with the scene changes they generate -- e.g., shadows, reflections, and motion of loose clothing. The layers can be individually retimed and recombined into a new video, allowing us to achieve realistic, high-quality renderings of retiming effects for real-world videos depicting complex actions and involving multiple individuals, including dancing, trampoline jumping, or group running.