Single-Shot Freestyle Dance Reenactment

TL;DR

This paper introduces a novel three-network approach for single-image dance reenactment that produces realistic, natural motion animations of diverse body types from unseen video sequences.

Contribution

It presents a new method combining segmentation, rendering, and face refinement networks to achieve high-quality, realistic dance reenactment from a single image.

Findings

Outperforms previous methods in visual quality

Capable of animating diverse body types and appearances

Handles challenging poses effectively

Abstract

The task of motion transfer between a source dancer and a target person is a special case of the pose transfer problem, in which the target person changes their pose in accordance with the motions of the dancer. In this work, we propose a novel method that can reanimate a single image by arbitrary video sequences, unseen during training. The method combines three networks: (i) a segmentation-mapping network, (ii) a realistic frame-rendering network, and (iii) a face refinement network. By separating this task into three stages, we are able to attain a novel sequence of realistic frames, capturing natural motion and appearance. Our method obtains significantly better visual quality than previous methods and is able to animate diverse body types and appearances, which are captured in challenging poses, as shown in the experiments and supplementary video.