ハミングによる検索機能を備えた音楽配信システムの開発

Music retrieval systems are extremely useful for collecting digital music data from
on-line music distribution sites. Especially, there is a great need to develop effective
techniques for content-based music retrieval systems, which can retrieve by humming
query. The main issues in this research is how to decide the similarity of each music
features extracted from music data. In order to calculate the similarity, some conventional
methods use Euclid distance or DP matching, but it is very hard to solve the problem of the
vagueness of humming query. In this paper, we propose a new similar music retrieval
method based on humming query using the Earth Mover's Distance as the distance measure.
Computing the EMD is based on a solution to the transportation problem, and the EMD is
applied as the distance measure on similar image retrieval systems. In addition, we focus
that the time complexity of the EMD is exponential worst case toward the number of
notes, the improved method to decrease the number of notes in the music feature is also
proposed. Experimental results show that the proposed method can improve the retrieval
precision of conventional systems.