Automatic retrieval of consistent dance motion and music through rhythmic structure

• Main Author: 謝偉鴻
• Other Authors: 林奕成
• Format: Others
• Language: en_US
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/53555736867908832933

碩士 === 國立交通大學 === 資訊科學與工程研究所 === 100 === Dance is an art form that often accompanies corresponding music with rhythm. To compose a dance with the accompanied music, the choreographers usually focus on rhythmic articulation, theme and variation, etc. The rhythms are essential for producing dancing motion and its accompanied music. In this thesis, we propose a method that searches best-matching music in database according to the input dance motion without any domain knowledge between music design and dance design. Based on the popular music (rock, pop, hip-hop, etc.) in our database, we employ constrained clustering algorithm to preprocess their cluster paths. We then rank a score through their corresponding rhythm between dance motion and music. This score focuses on their repeated rhythmic parts and consistency points. Our goal is to help the users retrieve best-fit pairs between dance motion and music with their rhythmic structure.