Trainable COPE Features for Sound Event Detection

Nicola Strisciuglio; Nicolai Petkov

doi:10.1007/978-3-030-33904-3_56

Trainable COPE Features for Sound Event Detection

Nicola Strisciuglio^*, Nicolai Petkov

^*Corresponding author for this work

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

1 Citation (Scopus)

Abstract

Systems for automatic analysis of sounds and detection of events are of great importance as they can be used as substitutes of or complement to video analytic systems. In this paper we describe a flexible system for the detection of audio events based on the use of trainable COPE (Combination of Peaks of Energy) features. The structure of a COPE feature is determined in an automatic configuration process on a single prototype example. Thus, they can be adapted to different kinds of sounds of interest. We configure a set of COPE features in order to account for robustness to variations of the characteristics of sounds within a specific class. The proposed system is flexible as new features (also configured on examples drawn from new classes) can be easily added to the feature set. We performed experiments on the MIVIA road events data set for road surveillance applications and compared the results that we achieved with the ones of other existing methods.

Original language	English
Title of host publication	Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications
Subtitle of host publication	24th Iberoamerican Congress
Editors	Ingela Nyström, Yanio Hernández Heredia, Vladimir Milián Núñez
Publisher	Springer
Pages	599-609
Number of pages	11
ISBN (Electronic)	978-3-030-33904-3
ISBN (Print)	978-3-030-33903-6
DOIs	https://doi.org/10.1007/978-3-030-33904-3_56
Publication status	Published - 1 Jan 2019
Externally published	Yes
Event	24th Iberoamerican Congress on Pattern Recognition, CIARP 2019 - Havana, Cuba Duration: 28 Oct 2019 → 31 Oct 2019 Conference number: 24

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11896 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	24th Iberoamerican Congress on Pattern Recognition, CIARP 2019
Abbreviated title	CIARP 2019
Country/Territory	Cuba
City	Havana
Period	28/10/19 → 31/10/19

Keywords

Few shot training
Sound event detection
Trainable features

Access to Document

10.1007/978-3-030-33904-3_56

Cite this

Strisciuglio, N., & Petkov, N. (2019). Trainable COPE Features for Sound Event Detection. In I. Nyström, Y. Hernández Heredia, & V. Milián Núñez (Eds.), Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress (pp. 599-609). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11896 LNCS). Springer. https://doi.org/10.1007/978-3-030-33904-3_56

Strisciuglio, Nicola ; Petkov, Nicolai. / Trainable COPE Features for Sound Event Detection. Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress. editor / Ingela Nyström ; Yanio Hernández Heredia ; Vladimir Milián Núñez. Springer, 2019. pp. 599-609 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{2f90b3ab84dd4bfdbf18b43308f37d1e,

title = "Trainable COPE Features for Sound Event Detection",

abstract = "Systems for automatic analysis of sounds and detection of events are of great importance as they can be used as substitutes of or complement to video analytic systems. In this paper we describe a flexible system for the detection of audio events based on the use of trainable COPE (Combination of Peaks of Energy) features. The structure of a COPE feature is determined in an automatic configuration process on a single prototype example. Thus, they can be adapted to different kinds of sounds of interest. We configure a set of COPE features in order to account for robustness to variations of the characteristics of sounds within a specific class. The proposed system is flexible as new features (also configured on examples drawn from new classes) can be easily added to the feature set. We performed experiments on the MIVIA road events data set for road surveillance applications and compared the results that we achieved with the ones of other existing methods.",

keywords = "Few shot training, Sound event detection, Trainable features",

author = "Nicola Strisciuglio and Nicolai Petkov",

year = "2019",

month = jan,

day = "1",

doi = "10.1007/978-3-030-33904-3_56",

language = "English",

isbn = "978-3-030-33903-6",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer",

pages = "599--609",

editor = "Ingela Nystr{\"o}m and {Hern{\'a}ndez Heredia}, Yanio and {Mili{\'a}n N{\'u}{\~n}ez}, Vladimir",

booktitle = "Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications",

address = "Germany",

note = "24th Iberoamerican Congress on Pattern Recognition, CIARP 2019, CIARP 2019 ; Conference date: 28-10-2019 Through 31-10-2019",

}

Strisciuglio, N & Petkov, N 2019, Trainable COPE Features for Sound Event Detection. in I Nyström, Y Hernández Heredia & V Milián Núñez (eds), Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11896 LNCS, Springer, pp. 599-609, 24th Iberoamerican Congress on Pattern Recognition, CIARP 2019, Havana, Cuba, 28/10/19. https://doi.org/10.1007/978-3-030-33904-3_56

Trainable COPE Features for Sound Event Detection. / Strisciuglio, Nicola; Petkov, Nicolai.
Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress. ed. / Ingela Nyström; Yanio Hernández Heredia; Vladimir Milián Núñez. Springer, 2019. p. 599-609 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11896 LNCS).

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

TY - GEN

T1 - Trainable COPE Features for Sound Event Detection

AU - Strisciuglio, Nicola

AU - Petkov, Nicolai

N1 - Conference code: 24

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Systems for automatic analysis of sounds and detection of events are of great importance as they can be used as substitutes of or complement to video analytic systems. In this paper we describe a flexible system for the detection of audio events based on the use of trainable COPE (Combination of Peaks of Energy) features. The structure of a COPE feature is determined in an automatic configuration process on a single prototype example. Thus, they can be adapted to different kinds of sounds of interest. We configure a set of COPE features in order to account for robustness to variations of the characteristics of sounds within a specific class. The proposed system is flexible as new features (also configured on examples drawn from new classes) can be easily added to the feature set. We performed experiments on the MIVIA road events data set for road surveillance applications and compared the results that we achieved with the ones of other existing methods.

AB - Systems for automatic analysis of sounds and detection of events are of great importance as they can be used as substitutes of or complement to video analytic systems. In this paper we describe a flexible system for the detection of audio events based on the use of trainable COPE (Combination of Peaks of Energy) features. The structure of a COPE feature is determined in an automatic configuration process on a single prototype example. Thus, they can be adapted to different kinds of sounds of interest. We configure a set of COPE features in order to account for robustness to variations of the characteristics of sounds within a specific class. The proposed system is flexible as new features (also configured on examples drawn from new classes) can be easily added to the feature set. We performed experiments on the MIVIA road events data set for road surveillance applications and compared the results that we achieved with the ones of other existing methods.

KW - Few shot training

KW - Sound event detection

KW - Trainable features

U2 - 10.1007/978-3-030-33904-3_56

DO - 10.1007/978-3-030-33904-3_56

M3 - Conference contribution

AN - SCOPUS:85075694192

SN - 978-3-030-33903-6

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 599

EP - 609

BT - Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications

A2 - Nyström, Ingela

A2 - Hernández Heredia, Yanio

A2 - Milián Núñez, Vladimir

PB - Springer

T2 - 24th Iberoamerican Congress on Pattern Recognition, CIARP 2019

Y2 - 28 October 2019 through 31 October 2019

ER -

Strisciuglio N, Petkov N. Trainable COPE Features for Sound Event Detection. In Nyström I, Hernández Heredia Y, Milián Núñez V, editors, Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications: 24th Iberoamerican Congress. Springer. 2019. p. 599-609. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-33904-3_56

Trainable COPE Features for Sound Event Detection

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this