Hierarchical Video Generation for Complex Data

TL;DR

This paper introduces a hierarchical, coarse-to-fine video generation model that efficiently creates high-resolution, multi-frame videos by sequentially training on partial views, scaling to complex datasets.

Contribution

A novel hierarchical model for video generation that reduces computational complexity and enables high-resolution, multi-frame video synthesis.

Findings

Successfully generates 256x256 videos with 48 frames

Scales to high-resolution videos beyond a few frames

Validated on Kinetics-600 and BDD100K datasets

Abstract

Videos can often be created by first outlining a global description of the scene and then adding local details. Inspired by this we propose a hierarchical model for video generation which follows a coarse to fine approach. First our model generates a low resolution video, establishing the global scene structure, that is then refined by subsequent levels in the hierarchy. We train each level in our hierarchy sequentially on partial views of the videos. This reduces the computational complexity of our generative model, which scales to high-resolution videos beyond a few frames. We validate our approach on Kinetics-600 and BDD100K, for which we train a three level model capable of generating 256x256 videos with 48 frames.