Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge

Christian Kroos; Oliver Bones; Yin Cao; Lara Harris; Philip J. B.  Jackson; William J.  Davies; Wenwu Wang; Trevor J. Cox; Mark D. Plumbley

doi:10.1109/icassp.2019.8683292

Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge

Christian Kroos, Oliver Bones, Yin Cao, Lara Harris, Philip J. B. Jackson, William J. Davies, Wenwu Wang, Trevor J. Cox, Mark D. Plumbley

Centre for Manufacturing and Materials

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

10 Citations (Scopus)

Abstract

Humans are able to identify a large number of environmental sounds and categorise them according to high-level semantic categories, e.g. urban sounds or music. They are also capable of generalising from past experience to new sounds when applying these categories. In this paper we report on the creation of a data set that is structured according to the top-level of a taxonomy derived from human judgements and the design of an associated machine learning challenge, in which strong generalisation abilities are required to be successful. We introduce a baseline classification system, a deep convolutional network, which showed strong performance with an average accuracy on the evaluation data of 80.8%. The result is discussed in the light of two alternative explanations: An unlikely accidental category bias in the sound recordings or a more plausible true acoustic grounding of the high-level categories.

Original language	English
Title of host publication	ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publisher	IEEE
Pages	8082-8086
Number of pages	5
ISBN (Electronic)	978-1-4799-8131-1
ISBN (Print)	978-1-4799-8132-8
DOIs	https://doi.org/10.1109/icassp.2019.8683292
Publication status	E-pub ahead of print - 17 Apr 2019
Event	44th International Conference on Acoustics, Speech, and Signal Processing - Brighton, United Kingdom Duration: 12 May 2019 → 17 May 2019 https://www.2019.ieeeicassp.org/

Publication series

Name
Publisher	IEEE
ISSN (Print)	1520-6149
ISSN (Electronic)	2379-190X

Conference

Conference	44th International Conference on Acoustics, Speech, and Signal Processing
Abbreviated title	ICASSP 2019
Country/Territory	United Kingdom
City	Brighton
Period	12/05/19 → 17/05/19
Internet address	https://www.2019.ieeeicassp.org/

Keywords

acoustic classification
machine learning challenge
sound taxonomy
deep learning
convolutional neural network

Access to Document

10.1109/icassp.2019.8683292

Cite this

Kroos, C., Bones, O., Cao, Y., Harris, L., Jackson, P. J. B., Davies, W. J., Wang, W., Cox, T. J., & Plumbley, M. D. (2019). Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge. In ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) (pp. 8082-8086). IEEE. Advance online publication. https://doi.org/10.1109/icassp.2019.8683292

Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge. / Kroos, Christian; Bones, Oliver; Cao, Yin et al.
ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2019. p. 8082-8086.

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

Kroos, C, Bones, O, Cao, Y, Harris, L, Jackson, PJB, Davies, WJ, Wang, W, Cox, TJ & Plumbley, MD 2019, Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge. in ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, pp. 8082-8086, 44th International Conference on Acoustics, Speech, and Signal Processing, Brighton, United Kingdom, 12/05/19. https://doi.org/10.1109/icassp.2019.8683292

@inproceedings{bb2d06199309486b80ccd26f1bca1884,

title = "Generalisation in Environmental Sound Classification: The {\textquoteleft}Making Sense of Sounds{\textquoteright} Data Set and Challenge",

abstract = "Humans are able to identify a large number of environmental sounds and categorise them according to high-level semantic categories, e.g. urban sounds or music. They are also capable of generalising from past experience to new sounds when applying these categories. In this paper we report on the creation of a data set that is structured according to the top-level of a taxonomy derived from human judgements and the design of an associated machine learning challenge, in which strong generalisation abilities are required to be successful. We introduce a baseline classification system, a deep convolutional network, which showed strong performance with an average accuracy on the evaluation data of 80.8%. The result is discussed in the light of two alternative explanations: An unlikely accidental category bias in the sound recordings or a more plausible true acoustic grounding of the high-level categories.",

keywords = "acoustic classification, machine learning challenge, sound taxonomy, deep learning, convolutional neural network",

author = "Christian Kroos and Oliver Bones and Yin Cao and Lara Harris and Jackson, {Philip J. B.} and Davies, {William J.} and Wenwu Wang and Cox, {Trevor J.} and Plumbley, {Mark D.}",

year = "2019",

month = apr,

day = "17",

doi = "10.1109/icassp.2019.8683292",

language = "English",

isbn = "978-1-4799-8132-8",

publisher = "IEEE",

pages = "8082--8086",

booktitle = "ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",

address = "United States",

note = "44th International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 ; Conference date: 12-05-2019 Through 17-05-2019",

url = "https://www.2019.ieeeicassp.org/",

}

TY - GEN

T1 - Generalisation in Environmental Sound Classification

T2 - 44th International Conference on Acoustics, Speech, and Signal Processing

AU - Kroos, Christian

AU - Bones, Oliver

AU - Cao, Yin

AU - Harris, Lara

AU - Jackson, Philip J. B.

AU - Davies, William J.

AU - Wang, Wenwu

AU - Cox, Trevor J.

AU - Plumbley, Mark D.

PY - 2019/4/17

Y1 - 2019/4/17

N2 - Humans are able to identify a large number of environmental sounds and categorise them according to high-level semantic categories, e.g. urban sounds or music. They are also capable of generalising from past experience to new sounds when applying these categories. In this paper we report on the creation of a data set that is structured according to the top-level of a taxonomy derived from human judgements and the design of an associated machine learning challenge, in which strong generalisation abilities are required to be successful. We introduce a baseline classification system, a deep convolutional network, which showed strong performance with an average accuracy on the evaluation data of 80.8%. The result is discussed in the light of two alternative explanations: An unlikely accidental category bias in the sound recordings or a more plausible true acoustic grounding of the high-level categories.

AB - Humans are able to identify a large number of environmental sounds and categorise them according to high-level semantic categories, e.g. urban sounds or music. They are also capable of generalising from past experience to new sounds when applying these categories. In this paper we report on the creation of a data set that is structured according to the top-level of a taxonomy derived from human judgements and the design of an associated machine learning challenge, in which strong generalisation abilities are required to be successful. We introduce a baseline classification system, a deep convolutional network, which showed strong performance with an average accuracy on the evaluation data of 80.8%. The result is discussed in the light of two alternative explanations: An unlikely accidental category bias in the sound recordings or a more plausible true acoustic grounding of the high-level categories.

KW - acoustic classification

KW - machine learning challenge

KW - sound taxonomy

KW - deep learning

KW - convolutional neural network

U2 - 10.1109/icassp.2019.8683292

DO - 10.1109/icassp.2019.8683292

M3 - Conference proceeding

SN - 978-1-4799-8132-8

SP - 8082

EP - 8086

BT - ICASSP 2019 - 2019 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

PB - IEEE

Y2 - 12 May 2019 through 17 May 2019

ER -

Generalisation in Environmental Sound Classification: The ‘Making Sense of Sounds’ Data Set and Challenge

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this