Acoustic Scene Classification Using A Convolutional Neural Network Ensemble and Nearest Neighbor Filters

Research output: Contribution to conferencePosterResearchpeer-review

Abstract

This paper proposes Convolutional Neural Network (CNN) ensembles for acoustic scene classification of tasks 1A and 1B of the DCASE 2018 challenge. We introduce a nearest neighbor filter applied on spectrograms, which allows to emphasize and smooth similar patterns of sound events in a scene. We also propose a variety of CNN models for single-input (SI) and multi-input (MI) channels and three different methods for building a network ensemble. The experimental results show that for task 1A the combination of the MI-CNN structures using both of log-mel features and their nearest neighbor filtering is slightly more effective than the single-input channel CNN models using log-mel features only. This statement is opposite for task 1B. In addition, the ensemble methods improve the accuracy of the system significantly, the best ensemble method is ensemble selection, which achieves 69.3% for task 1A and 63.6% for task 1B. This improves the baseline system by 8.9% and 14.4% for task 1A and 1B, respectively.
LanguageEnglish
StatusPublished - 20 Nov 2018

Fingerprint

Acoustics
Neural networks
Acoustic waves

ASJC Scopus subject areas

  • Engineering(all)

Cite this

@conference{1911737d0c0f45e88d878023b080f63f,
title = "Acoustic Scene Classification Using A Convolutional Neural Network Ensemble and Nearest Neighbor Filters",
abstract = "This paper proposes Convolutional Neural Network (CNN) ensembles for acoustic scene classification of tasks 1A and 1B of the DCASE 2018 challenge. We introduce a nearest neighbor filter applied on spectrograms, which allows to emphasize and smooth similar patterns of sound events in a scene. We also propose a variety of CNN models for single-input (SI) and multi-input (MI) channels and three different methods for building a network ensemble. The experimental results show that for task 1A the combination of the MI-CNN structures using both of log-mel features and their nearest neighbor filtering is slightly more effective than the single-input channel CNN models using log-mel features only. This statement is opposite for task 1B. In addition, the ensemble methods improve the accuracy of the system significantly, the best ensemble method is ensemble selection, which achieves 69.3{\%} for task 1A and 63.6{\%} for task 1B. This improves the baseline system by 8.9{\%} and 14.4{\%} for task 1A and 1B, respectively.",
author = "Nguyen, {Thi Kim Truc} and Franz Pernkopf",
year = "2018",
month = "11",
day = "20",
language = "English",

}

TY - CONF

T1 - Acoustic Scene Classification Using A Convolutional Neural Network Ensemble and Nearest Neighbor Filters

AU - Nguyen, Thi Kim Truc

AU - Pernkopf, Franz

PY - 2018/11/20

Y1 - 2018/11/20

N2 - This paper proposes Convolutional Neural Network (CNN) ensembles for acoustic scene classification of tasks 1A and 1B of the DCASE 2018 challenge. We introduce a nearest neighbor filter applied on spectrograms, which allows to emphasize and smooth similar patterns of sound events in a scene. We also propose a variety of CNN models for single-input (SI) and multi-input (MI) channels and three different methods for building a network ensemble. The experimental results show that for task 1A the combination of the MI-CNN structures using both of log-mel features and their nearest neighbor filtering is slightly more effective than the single-input channel CNN models using log-mel features only. This statement is opposite for task 1B. In addition, the ensemble methods improve the accuracy of the system significantly, the best ensemble method is ensemble selection, which achieves 69.3% for task 1A and 63.6% for task 1B. This improves the baseline system by 8.9% and 14.4% for task 1A and 1B, respectively.

AB - This paper proposes Convolutional Neural Network (CNN) ensembles for acoustic scene classification of tasks 1A and 1B of the DCASE 2018 challenge. We introduce a nearest neighbor filter applied on spectrograms, which allows to emphasize and smooth similar patterns of sound events in a scene. We also propose a variety of CNN models for single-input (SI) and multi-input (MI) channels and three different methods for building a network ensemble. The experimental results show that for task 1A the combination of the MI-CNN structures using both of log-mel features and their nearest neighbor filtering is slightly more effective than the single-input channel CNN models using log-mel features only. This statement is opposite for task 1B. In addition, the ensemble methods improve the accuracy of the system significantly, the best ensemble method is ensemble selection, which achieves 69.3% for task 1A and 63.6% for task 1B. This improves the baseline system by 8.9% and 14.4% for task 1A and 1B, respectively.

M3 - Poster

ER -