Listen to Rhythm, Choose Movements: Autoregressive Multimodal Dance Generation via Diffusion and Mamba with Decoupled Dance Dataset

TL;DR

This paper introduces LRCM, a multimodal-guided diffusion framework for dance motion generation that achieves coherent, long-duration sequences by integrating audio, text, and motion data with a novel architecture.

Contribution

The work presents a new decoupling paradigm for dance datasets and a diffusion architecture with a Motion Temporal Mamba Module for improved long-sequence dance synthesis.

Findings

LRCM outperforms existing methods in quantitative metrics.

The framework supports diverse multimodal inputs.

LRCM generates smooth, long-duration dance sequences.

Abstract

Advances in generative models and sequence learning have greatly promoted research in dance motion generation, yet current methods still suffer from coarse semantic control and poor coherence in long sequences. In this work, we present Listen to Rhythm, Choose Movements (LRCM), a multimodal-guided diffusion framework supporting both diverse input modalities and autoregressive dance motion generation. We explore a feature decoupling paradigm for dance datasets and generalize it to the Motorica Dance dataset, separating motion capture data, audio rhythm, and professionally annotated global and local text descriptions. Our diffusion architecture integrates an audio-latent Conformer and a text-latent Cross-Conformer, and incorporates a Motion Temporal Mamba Module (MTMM) to enable smooth, long-duration autoregressive synthesis. Experimental results indicate that LRCM delivers strong performance in both functional capability and quantitative metrics, demonstrating notable potential in multimodal input scenarios and extended sequence generation. The project page is available at https://oranduanstudy.github.io/LRCM/.