ST-GDance: Long-Term and Collision-Free Group Choreography from Music

TL;DR

ST-GDance is a new framework that efficiently generates long-term, collision-free group dance choreography from music by decoupling spatial and temporal modeling, outperforming existing methods.

Contribution

It introduces a novel decoupled spatial-temporal approach with lightweight graph convolutions and sparse attention for scalable, collision-free dance generation.

Findings

Outperforms state-of-the-art baselines on the AIOZ-GDance dataset.

Effectively models dense spatial-temporal interactions for long sequences.

Reduces computational costs while maintaining smooth, collision-free dances.

Abstract

Group dance generation from music has broad applications in film, gaming, and animation production. However, it requires synchronizing multiple dancers while maintaining spatial coordination. As the number of dancers and sequence length increase, this task faces higher computational complexity and a greater risk of motion collisions. Existing methods often struggle to model dense spatial-temporal interactions, leading to scalability issues and multi-dancer collisions. To address these challenges, we propose ST-GDance, a novel framework that decouples spatial and temporal dependencies to optimize long-term and collision-free group choreography. We employ lightweight graph convolutions for distance-aware spatial modeling and accelerated sparse attention for efficient temporal modeling. This design significantly reduces computational costs while ensuring smooth and collision-free interactions. Experiments on the AIOZ-GDance dataset demonstrate that ST-GDance outperforms state-of-the-art baselines, particularly in generating long and coherent group dance sequences. Project page: https://yilliajing.github.io/ST-GDance-Website/.