GCDance: Genre-Controlled 3D Full Body Dance Generation Driven By Music

Xinran Liu, Xu Dong, Diptesh Kanojia, Wenwu Wang, Zhenhua Feng · 2025

Read GCDance: Genre-Controlled 3D Full Body Dance Generation Driven By Music by Xinran Liu, Xu Dong, Diptesh Kanojia, Wenwu Wang, Zhenhua Feng (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2502-18309.txt.

Description

Xinran Liu and four co-authors present, in a 2025 paper, a system that generates full-body dance in 3D, steered by both music and a text label for the genre. A good routine must honour the style of its genre while landing moves on the beat, and most earlier generators blur the style. Their framework, GCDance, is built on a diffusion model.

A text control turns a genre name or a free-form description into a control signal that guides the motion. Features from a music foundation model keep the dance and the words in step. A multi-task training rule balances realism, accuracy in space, and genre fit. On two public dance sets, the system beat the prior state of the art.

Catalog record

CREATORXinran Liu, Xu Dong, Diptesh Kanojia, Wenwu Wang, Zhenhua Feng
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE39 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2502-18309

Browse the full library · About the Tower of Babel