Please use this identifier to cite or link to this item:
Title: Building a large-scale dataset for audio-conditioned dance motion synthesis
Authors: Wu, Jinyi
Keywords: Engineering::Computer science and engineering::Computing methodologies::Image processing and computer vision
Issue Date: 2022
Publisher: Nanyang Technological University
Source: Wu, J. (2022). Building a large-scale dataset for audio-conditioned dance motion synthesis. Master's thesis, Nanyang Technological University, Singapore.
Abstract: 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.
DOI: 10.32657/10356/160410
Schools: School of Computer Science and Engineering 
Rights: This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License (CC BY-NC 4.0).
Fulltext Permission: open
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Theses

Files in This Item:
File Description SizeFormat 
Master_Thesis_Jinyi.pdf23.98 MBAdobe PDFThumbnail

Page view(s)

Updated on Sep 27, 2023

Download(s) 50

Updated on Sep 27, 2023

Google ScholarTM




Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.