An Adaptation Framework for QBH-based Music Retrieval

Springer Lecture Notes on Computer Science (LNCS), 11th International Conference on Knowledge-Based and Intelligent Information & Engineering Systems (KES 2007)

Seungmin Rho(1), Byeong-jun Han(2), Eenjun Hwang(2) Minkoo Kim(1),

(1) Graduate School of Information and Communication, Artificial Intelligence Laboratory.
(2) School of Electrical Engineering, Korea University.

Abstract

In this paper, we present a new music query transcription and refinement scheme for efficient music retrieval. For the accurate music query transcription into symbolic representation, we propose a method called WAE for note onset detection, and DTC for ADF onset detection. Also, in order to improve the retrieval performance, we propose a new relevance feedback scheme using genetic algorithm. We have built a prototype system based on this scheme and performed various experiments. Experimental results show that our proposed scheme achieves a good performance.

Full paper

Available on SpringerLink.
Linked on DBLP.

Presentation

Not Available Now.

Last Updated: 2007-09-28