A robust approach to content-based musical genre classification and retrieval using multi-feature clustering

Abstract

In this paper, we propose a new robust content-based musical genre classification and retrieval algorithm using multi-feature clustering (MFC) method. In contrast to previous works, this paper focuses on two practical issues of the system dependency problem on different input query patterns (or portions) and input query lengths which causes serious uncertainty of the system performance. In order to solve these problems, a new approach called multi-feature clustering (MFC) based on k-means clustering is proposed. To verify the performance of the proposed method, several excerpts with variable duration were extracted from every other position in a queried music file. Effectiveness of the system with MFC and without MFC is compared in terms of the classification and retrieval accuracy. It is demonstrated that the use of MFC significantly improves the system stability of musical genre classification and retrieval performance with higher accuracy rate. © Springer-Verlag Berlin Heidelberg 2004.