Building a large-scale dataset for audio-conditioned dance motion synthesis

• Main Author: Wu, Jinyi
• Other Authors: Chen Change Loy
• Format: Thesis-Master by Research
• Language: English
• Published: Nanyang Technological University 2022
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
• Online Access: https://hdl.handle.net/10356/160410
• Institution: Nanyang Technological University

Generative models for audio-conditioned dance motion synthesis map music features to dance movements. Models are trained with a few assumptions such as strong music-dance correlation, controlled motion data and relatively simple poses. These characteristics are found in all existing datasets for dance motion synthesis, and indeed recent methods can achieve good results. We introduce a new dataset aiming to challenge these common assumptions. We focus on breakdancing which features acrobatic moves and tangled postures. We source our data from the Red Bull BC One competition videos and adopt a hybrid labelling pipeline leveraging deep estimation models as well as manual annotations to obtain good quality keypoint sequences at a reduced cost. Our dataset can readily foster advance in dance motion synthesis. With intri- cate poses and swift movements, models are forced to go beyond learning a mapping between modalities and reason more effectively about body structure and movements.