Unsupervised feature learning for urban sound classification

Justin Salamon, Juan Pablo Bello

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Recent studies have demonstrated the potential of unsupervised feature learning for sound classification. In this paper we further explore the application of the spherical k-means algorithm for feature learning from audio signals, here in the domain of urban sound classification. Spherical k-means is a relatively simple technique that has recently been shown to be competitive with other more complex and time consuming approaches. We study how different parts of the processing pipeline influence performance, taking into account the specificities of the urban sonic environment. We evaluate our approach on the largest public dataset of urban sound sources available for research, and compare it to a baseline system based on MFCCs. We show that feature learning can outperform the baseline approach by configuring it to capture the temporal dynamics of urban sources. The results are complemented with error analysis and some proposals for future research.

Original languageEnglish (US)
Title of host publication2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages171-175
Number of pages5
Volume2015-August
ISBN (Electronic)9781467369978
DOIs
StatePublished - Aug 4 2015
Event40th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Brisbane, Australia
Duration: Apr 19 2014Apr 24 2014

Other

Other40th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015
CountryAustralia
CityBrisbane
Period4/19/144/24/14

Fingerprint

Acoustic waves
Error analysis
Pipelines
Processing

Keywords

  • machine learning
  • sound classification
  • spherical k-means
  • Unsupervised learning
  • urban

ASJC Scopus subject areas

  • Signal Processing
  • Software
  • Electrical and Electronic Engineering

Cite this

Salamon, J., & Bello, J. P. (2015). Unsupervised feature learning for urban sound classification. In 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings (Vol. 2015-August, pp. 171-175). [7177954] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2015.7177954

Unsupervised feature learning for urban sound classification. / Salamon, Justin; Bello, Juan Pablo.

2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. Vol. 2015-August Institute of Electrical and Electronics Engineers Inc., 2015. p. 171-175 7177954.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Salamon, J & Bello, JP 2015, Unsupervised feature learning for urban sound classification. in 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. vol. 2015-August, 7177954, Institute of Electrical and Electronics Engineers Inc., pp. 171-175, 40th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015, Brisbane, Australia, 4/19/14. https://doi.org/10.1109/ICASSP.2015.7177954
Salamon J, Bello JP. Unsupervised feature learning for urban sound classification. In 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. Vol. 2015-August. Institute of Electrical and Electronics Engineers Inc. 2015. p. 171-175. 7177954 https://doi.org/10.1109/ICASSP.2015.7177954
Salamon, Justin ; Bello, Juan Pablo. / Unsupervised feature learning for urban sound classification. 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings. Vol. 2015-August Institute of Electrical and Electronics Engineers Inc., 2015. pp. 171-175
@inproceedings{d0fb6e6b8beb49479221d53bd456bed2,
title = "Unsupervised feature learning for urban sound classification",
abstract = "Recent studies have demonstrated the potential of unsupervised feature learning for sound classification. In this paper we further explore the application of the spherical k-means algorithm for feature learning from audio signals, here in the domain of urban sound classification. Spherical k-means is a relatively simple technique that has recently been shown to be competitive with other more complex and time consuming approaches. We study how different parts of the processing pipeline influence performance, taking into account the specificities of the urban sonic environment. We evaluate our approach on the largest public dataset of urban sound sources available for research, and compare it to a baseline system based on MFCCs. We show that feature learning can outperform the baseline approach by configuring it to capture the temporal dynamics of urban sources. The results are complemented with error analysis and some proposals for future research.",
keywords = "machine learning, sound classification, spherical k-means, Unsupervised learning, urban",
author = "Justin Salamon and Bello, {Juan Pablo}",
year = "2015",
month = "8",
day = "4",
doi = "10.1109/ICASSP.2015.7177954",
language = "English (US)",
volume = "2015-August",
pages = "171--175",
booktitle = "2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
address = "United States",

}

TY - GEN

T1 - Unsupervised feature learning for urban sound classification

AU - Salamon, Justin

AU - Bello, Juan Pablo

PY - 2015/8/4

Y1 - 2015/8/4

N2 - Recent studies have demonstrated the potential of unsupervised feature learning for sound classification. In this paper we further explore the application of the spherical k-means algorithm for feature learning from audio signals, here in the domain of urban sound classification. Spherical k-means is a relatively simple technique that has recently been shown to be competitive with other more complex and time consuming approaches. We study how different parts of the processing pipeline influence performance, taking into account the specificities of the urban sonic environment. We evaluate our approach on the largest public dataset of urban sound sources available for research, and compare it to a baseline system based on MFCCs. We show that feature learning can outperform the baseline approach by configuring it to capture the temporal dynamics of urban sources. The results are complemented with error analysis and some proposals for future research.

AB - Recent studies have demonstrated the potential of unsupervised feature learning for sound classification. In this paper we further explore the application of the spherical k-means algorithm for feature learning from audio signals, here in the domain of urban sound classification. Spherical k-means is a relatively simple technique that has recently been shown to be competitive with other more complex and time consuming approaches. We study how different parts of the processing pipeline influence performance, taking into account the specificities of the urban sonic environment. We evaluate our approach on the largest public dataset of urban sound sources available for research, and compare it to a baseline system based on MFCCs. We show that feature learning can outperform the baseline approach by configuring it to capture the temporal dynamics of urban sources. The results are complemented with error analysis and some proposals for future research.

KW - machine learning

KW - sound classification

KW - spherical k-means

KW - Unsupervised learning

KW - urban

UR - http://www.scopus.com/inward/record.url?scp=84946051287&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84946051287&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2015.7177954

DO - 10.1109/ICASSP.2015.7177954

M3 - Conference contribution

AN - SCOPUS:84946051287

VL - 2015-August

SP - 171

EP - 175

BT - 2015 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2015 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -