BodyFormer: Semantics-guided 3D Body Gesture Synthesis with Transformer

TL;DR

This paper introduces BodyFormer, a transformer-based model that synthesizes diverse and realistic 3D body gestures from speech by modeling stochasticity and motion speed variations, even with limited data.

Contribution

The paper presents a variational transformer with mode positional embedding and intra-modal pre-training for speech-to-gesture synthesis, addressing data scarcity and gesture diversity.

Findings

Produces more realistic and appropriate gestures than state-of-the-art methods.

Effectively models stochastic gesture variations during speech.

Handles limited training data through pre-training scheme.

Abstract

Automatic gesture synthesis from speech is a topic that has attracted researchers for applications in remote communication, video games and Metaverse. Learning the mapping between speech and 3D full-body gestures is difficult due to the stochastic nature of the problem and the lack of a rich cross-modal dataset that is needed for training. In this paper, we propose a novel transformer-based framework for automatic 3D body gesture synthesis from speech. To learn the stochastic nature of the body gesture during speech, we propose a variational transformer to effectively model a probabilistic distribution over gestures, which can produce diverse gestures during inference. Furthermore, we introduce a mode positional embedding layer to capture the different motion speeds in different speaking modes. To cope with the scarcity of data, we design an intra-modal pre-training scheme that can learn the complex mapping between the speech and the 3D gesture from a limited amount of data. Our system is trained with either the Trinity speech-gesture dataset or the Talking With Hands 16.2M dataset. The results show that our system can produce more realistic, appropriate, and diverse body gestures compared to existing state-of-the-art approaches.