首页 | 本学科首页   官方微博 | 高级检索  
     


Stream segregation algorithm for pattern matching in polyphonic music databases
Authors:Wai Man Szeto  Man Hon Wong
Affiliation:(1) Department of Computer Science and Engineering, The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong SAR, China
Abstract:As music can be represented symbolically, most of the existing methods extend some string matching algorithms to retrieve musical patterns in a music database. However, not all retrieved patterns are perceptually significant because some of them are, in fact, inaudible. Music is perceived in groupings of musical notes called streams. The process of grouping musical notes into streams is called stream segregation. Stream-crossing musical patterns are perceptually insignificant and should be pruned from the retrieval results. This can be done if all musical notes in a music database are segregated into streams and musical patterns are retrieved from the streams. Findings in auditory psychology are utilized in this paper, in which stream segregation is modelled as a clustering process and an adapted single-link clustering algorithm is proposed. Supported by experiments on real music data, streams are identified by the proposed algorithm with considerable accuracy.
Contact InformationMan Hon WongEmail:
Keywords:Multimedia databases  Music information retrieval  Clustering algorithms
本文献已被 SpringerLink 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号