The complete code for training and testing, along with the dataset, will be publicly available.
Humans can perform a variety of interactive motions, among which two-person dance is one of the most challenging interactions. However, in terms of computer motion generation, current work is still unable to generate high-quality interactive motion, especially in the field of duet dance. On the one hand, this is caused by the lack of large-scale high-quality datasets. On the other hand, it arises from the incomplete representation of interactive motion and the lack of fine-grained optimization of interactions. To address these challenges, we propose a duet dance dataset that significantly enhances motion quality, data scale, and the variety of dance genres. Based on this dataset, we propose a new motion representation that can accurately and comprehensively describe interactive motion. We further introduce a diffusion-based algorithm with an interaction refine guidance strategy to optimize the realism of interactions progressively. Experiments demonstrate the effectiveness of our dataset and algorithm.
Rumba
|
Cha Cha
|
Waltz
|
Jazz
|
Shenyun
|
Jive
|
Tai
|
Uighur
|
Hantang
|
Hiphop
|
Dunhuang
|
Kpop
|
Urban
|
Miao
|
Samba
|
Rumba (React)
|
Waltz (React)
|
Sumba (React)
|
Shenyun (React)
|
Samba (React)
|
Waltz (React)
|
Rumba (React)
|
Urban (React)
|
ShenYun (React)
|
Waltz (React)
|
Sumba (React)
|
Waltz (React)
|
Waltz (duet)
|
Rumba (duet)
|
Chacha (duet)
|
Waltz (duet)
|
Waltz (duet)
|
Rumba (duet)
|
Waltz (duet)
|
Waltz (duet)
|
Urban (long)
|
Waltz (long)
|
Waltz (repaint)
|
Waltz (repaint)
|
Rumba (repaint)
|
Rumba (repaint)
|
Waltz (w/o leader contact)
|
Waltz (w/o leader contact)
|
Rumba (w/o leader contact)
|
Rumba (w/o leader contact)
|