A new methodology for music retrieval based on dynamic neural networks

  • Authors:

  • Laura E. Gomez;Humberto Sossa;Ricardo Barron;Julio F. Jimenez

  • Affiliations:

  • Centro de Investigación en Computación-IPN. Av. Juan de Dios Batiz S/N. Col. Nueva Industrial Vallejo, C.P. 07738, México, D.F.;Centro de Investigación en Computación-IPN. Av. Juan de Dios Batiz S/N. Col. Nueva Industrial Vallejo, C.P. 07738, México, D.F.;Centro de Investigación en Computación-IPN. Av. Juan de Dios Batiz S/N. Col. Nueva Industrial Vallejo, C.P. 07738, México, D.F.;Centro de Investigación en Computación-IPN. Av. Juan de Dios Batiz S/N. Col. Nueva Industrial Vallejo, C.P. 07738, México, D.F.

  • Venue:
  • International Journal of Hybrid Intelligent Systems
  • Year:
  • 2012

Quantified Score

Hi-index 0.00

Visualization

Abstract

Most of in music information retrieval MIR has been focused on the symbolic representations of music. However, most of the digitally available music is in the form of raw audio signals. Although various attempts for monophonic and polyphonic transcription have been developed, none has been successful and general enough in the case of real world signals. So far, most of the research has been based on developing efficient music retrieval systems. In this paper, we introduce a music retrieval system based on Dynamic Neural Networks DNN, which are trained with the signal melody, and not with traditional descriptors. The proposal was tested with a database composed of 1000 melodies. The results are very encouraging.