DNN transfer learning based non-linear feature extraction for acoustic event classification

Seongkyu Mun; Minkyu Shin; Suwon Shon; Wooil Kim; David K. Han; Hanseok Ko

doi:10.1587/transinf.2017EDL8048

DNN transfer learning based non-linear feature extraction for acoustic event classification

Seongkyu Mun, Minkyu Shin, Suwon Shon, Wooil Kim, David K. Han, Hanseok Ko

School of Electrical Engineering

Research output: Contribution to journal › Article › peer-review

12 Citations (Scopus)

Abstract

Recent acoustic event classification research has focused on training suitable filters to represent acoustic events. However, due to limited availability of target event databases and linearity of conventional filters, there is still room for improving performance. By exploiting the non-linear modeling of deep neural networks (DNNs) and their ability to learn beyond pre-Trained environments, this letter proposes a DNN-based feature extraction scheme for the classification of acoustic events. The effectiveness and robustness to noise of the proposed method are demonstrated using a database of indoor surveillance environments.

Original language	English
Pages (from-to)	2249-2252
Number of pages	4
Journal	IEICE Transactions on Information and Systems
Volume	E100D
Issue number	9
DOIs	https://doi.org/10.1587/transinf.2017EDL8048
Publication status	Published - 2017 Sept

Keywords

Acoustic event classification
Acoustic feature
Deep neural network
Transfer learning

ASJC Scopus subject areas

Software
Hardware and Architecture
Computer Vision and Pattern Recognition
Electrical and Electronic Engineering
Artificial Intelligence

Access to Document

10.1587/transinf.2017EDL8048

Cite this

@article{dab802441fea456b9691225836d9b036,

title = "DNN transfer learning based non-linear feature extraction for acoustic event classification",

abstract = "Recent acoustic event classification research has focused on training suitable filters to represent acoustic events. However, due to limited availability of target event databases and linearity of conventional filters, there is still room for improving performance. By exploiting the non-linear modeling of deep neural networks (DNNs) and their ability to learn beyond pre-Trained environments, this letter proposes a DNN-based feature extraction scheme for the classification of acoustic events. The effectiveness and robustness to noise of the proposed method are demonstrated using a database of indoor surveillance environments.",

keywords = "Acoustic event classification, Acoustic feature, Deep neural network, Transfer learning",

author = "Seongkyu Mun and Minkyu Shin and Suwon Shon and Wooil Kim and Han, {David K.} and Hanseok Ko",

note = "Funding Information: This material is based upon work supported by the Air Force Office of Scientific Research under award number FA2386-16-1-4130 and authors would like to thank the anonymous reviewers for their valuable comments. Publisher Copyright: Copyright {\textcopyright} 2017 The Institute of Electronics, Information and Communication Engineers.",

year = "2017",

month = sep,

doi = "10.1587/transinf.2017EDL8048",

language = "English",

volume = "E100D",

pages = "2249--2252",

journal = "IEICE Transactions on Information and Systems",

issn = "0916-8532",

publisher = "Maruzen Co., Ltd/Maruzen Kabushikikaisha",

number = "9",

}

TY - JOUR

T1 - DNN transfer learning based non-linear feature extraction for acoustic event classification

AU - Mun, Seongkyu

AU - Shin, Minkyu

AU - Shon, Suwon

AU - Kim, Wooil

AU - Han, David K.

AU - Ko, Hanseok

N1 - Funding Information: This material is based upon work supported by the Air Force Office of Scientific Research under award number FA2386-16-1-4130 and authors would like to thank the anonymous reviewers for their valuable comments. Publisher Copyright: Copyright © 2017 The Institute of Electronics, Information and Communication Engineers.

PY - 2017/9

Y1 - 2017/9

N2 - Recent acoustic event classification research has focused on training suitable filters to represent acoustic events. However, due to limited availability of target event databases and linearity of conventional filters, there is still room for improving performance. By exploiting the non-linear modeling of deep neural networks (DNNs) and their ability to learn beyond pre-Trained environments, this letter proposes a DNN-based feature extraction scheme for the classification of acoustic events. The effectiveness and robustness to noise of the proposed method are demonstrated using a database of indoor surveillance environments.

AB - Recent acoustic event classification research has focused on training suitable filters to represent acoustic events. However, due to limited availability of target event databases and linearity of conventional filters, there is still room for improving performance. By exploiting the non-linear modeling of deep neural networks (DNNs) and their ability to learn beyond pre-Trained environments, this letter proposes a DNN-based feature extraction scheme for the classification of acoustic events. The effectiveness and robustness to noise of the proposed method are demonstrated using a database of indoor surveillance environments.

KW - Acoustic event classification

KW - Acoustic feature

KW - Deep neural network

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=85029429098&partnerID=8YFLogxK

U2 - 10.1587/transinf.2017EDL8048

DO - 10.1587/transinf.2017EDL8048

M3 - Article

AN - SCOPUS:85029429098

SN - 0916-8532

VL - E100D

SP - 2249

EP - 2252

JO - IEICE Transactions on Information and Systems

JF - IEICE Transactions on Information and Systems

IS - 9

ER -

DNN transfer learning based non-linear feature extraction for acoustic event classification

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this