Compositional Video Synthesis with Action Graphs

TL;DR

This paper introduces a new method for generating videos conditioned on multiple coordinated actions using Action Graphs, enabling more complex and semantically consistent video synthesis with zero-shot capabilities.

Contribution

The paper proposes AG2Vid, a novel generative model that uses Action Graphs to synthesize videos with multiple simultaneous actions, improving visual quality and semantic coherence.

Findings

Better visual quality than baselines

Enhanced semantic consistency

Zero-shot action composition

Abstract

Videos of actions are complex signals containing rich compositional structure in space and time. Current video generation methods lack the ability to condition the generation on multiple coordinated and potentially simultaneous timed actions. To address this challenge, we propose to represent the actions in a graph structure called Action Graph and present the new ``Action Graph To Video'' synthesis task. Our generative model for this task (AG2Vid) disentangles motion and appearance features, and by incorporating a scheduling mechanism for actions facilitates a timely and coordinated video generation. We train and evaluate AG2Vid on the CATER and Something-Something V2 datasets, and show that the resulting videos have better visual quality and semantic consistency compared to baselines. Finally, our model demonstrates zero-shot abilities by synthesizing novel compositions of the learned actions. For code and pretrained models, see the project page https://roeiherz.github.io/AG2Video