Sound event detection for music signals using gaussian processes

  • Pablo A. Alvarado-Durán Universidad Tecnológica de Pereira, Pereira
  • Mauricio A. Álvarez-López Universidad Tecnológica de Pereira, Pereira
  • Álvaro A. Orozco-Gutiérrez Universidad Tecnológica de Pereira, Pereira
Keywords: Gaussian processes classification, supervised machine learning, spectrogram, event detection, music signals.

Abstract

In this paper we present a new methodology for detecting sound events in music signals using Gaussian Processes. Our method firstly takes a time-frequency representation, i.e. the spectrogram, of the input audio signal. Secondly the spectrogram dimension is reduced translating the linear Hertz frequency scale into the logarithmic Mel frequency scale using a triangular filter bank. Finally every short-time spectrum, i.e. every Mel spectrogram column, is classified as “Event” or “Not Event” by a Gaussian Processes Classifier. We compare our method with other event detection techniques widely used. To do so, we use MATLAB® to program each technique and test them using two datasets of music with different levels of complexity. Results show that the new methodology outperforms the standard approaches, getting an improvement by about 1.66 % on the dataset one and 0.45 % on the dataset two in terms of F-measure.

Author Biographies

Pablo A. Alvarado-Durán, Universidad Tecnológica de Pereira, Pereira
Ingeniero Electrónico, Programa de Ingeniería
Eléctrica, Universidad Tecnológica de Pereira,
Pereira
Mauricio A. Álvarez-López, Universidad Tecnológica de Pereira, Pereira
Doctor en Ciencias de la Computación,
Programa de Ingeniería Eléctrica,
Universidad Tecnológica de Pereira, Pereira
Álvaro A. Orozco-Gutiérrez, Universidad Tecnológica de Pereira, Pereira
Doctor en Bioingeniería, Programa de Ingeniería Eléctrica,
Universidad Tecnológica de Pereira, Pereira
How to Cite
Alvarado-Durán, P. A., Álvarez-López, M. A., & Orozco-Gutiérrez, Álvaro A. (2011). Sound event detection for music signals using gaussian processes. TecnoLógicas, (31), 93-122. https://doi.org/10.22430/22565337.108

Downloads

Download data is not yet available.
Published
2011-11-30
Section
Research Papers

Most read articles by the same author(s)