A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification

Tuan Nguyen; Dat Ngo; Lam Pham; Linh Tran; Trang Hoang

doi:10.1109/RIVF48685.2020.9140774

A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification

Tuan Nguyen, Dat Ngo, Lam Pham, Linh Tran, Trang Hoang

Data Science & Artifical Intelligence

Electrical and Electronics Department, Ho Chi Minh City University of Technology

Publikation: Beitrag in Buch oder Tagungsband › Vortrag mit Beitrag in Tagungsband › Begutachtung

Abstract

This paper proposes a deep learning framework applied for Acoustic Scene Classification (ASC), which identifies recording location. In general, we apply three types of spectrograms: Gammatone (GAM), log-Mel and Constant Q Transform (CQT) for front-end feature extraction. For back-end classification, we present a re-trained model with a multi-kernel CDNN-based architecture for the pre-trained process and a DNN-based network for the post-trained process. Our obtained results over DCASE 2016 dataset show a significant improvement, increasing by nearly 8% compared to DCASE baseline of 77.2%.

Originalsprache	Englisch
Titel	RIVF International Conference on Computing and Communication Technologies (RIVF) 2020
Seiten	1-5
Seitenumfang	5
DOIs	https://doi.org/10.1109/RIVF48685.2020.9140774
Publikationsstatus	Veröffentlicht - 2020
Veranstaltung	2020 RIVF International Conference on Computing and Communication Technologies (RIVF) - Dauer: 14 Okt. 2020 → 15 Okt. 2020

Konferenz

Konferenz	2020 RIVF International Conference on Computing and Communication Technologies (RIVF)
Zeitraum	14/10/20 → 15/10/20

Research Field

Ehemaliges Research Field - Data Science

Zugriff auf Dokument

10.1109/RIVF48685.2020.9140774

Diese Publikation zitieren

@inproceedings{fae3c15cfcea4c59988b9e00fd4e0b41,

title = "A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification",

abstract = "This paper proposes a deep learning framework applied for Acoustic Scene Classification (ASC), which identifies recording location. In general, we apply three types of spectrograms: Gammatone (GAM), log-Mel and Constant Q Transform (CQT) for front-end feature extraction. For back-end classification, we present a re-trained model with a multi-kernel CDNN-based architecture for the pre-trained process and a DNN-based network for the post-trained process. Our obtained results over DCASE 2016 dataset show a significant improvement, increasing by nearly 8% compared to DCASE baseline of 77.2%.",

author = "Tuan Nguyen and Dat Ngo and Lam Pham and Linh Tran and Trang Hoang",

year = "2020",

doi = "10.1109/RIVF48685.2020.9140774",

language = "English",

isbn = "978-1-7281-5377-3 3",

pages = "1--5",

booktitle = "RIVF International Conference on Computing and Communication Technologies (RIVF) 2020",

note = "2020 RIVF International Conference on Computing and Communication Technologies (RIVF) ; Conference date: 14-10-2020 Through 15-10-2020",

}

Nguyen, T, Ngo, D, Pham, L, Tran, L & Hoang, T 2020, A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification. in RIVF International Conference on Computing and Communication Technologies (RIVF) 2020. S. 1-5, 2020 RIVF International Conference on Computing and Communication Technologies (RIVF), 14/10/20. https://doi.org/10.1109/RIVF48685.2020.9140774

TY - GEN

T1 - A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification

AU - Nguyen, Tuan

AU - Ngo, Dat

AU - Pham, Lam

AU - Tran, Linh

AU - Hoang, Trang

PY - 2020

Y1 - 2020

N2 - This paper proposes a deep learning framework applied for Acoustic Scene Classification (ASC), which identifies recording location. In general, we apply three types of spectrograms: Gammatone (GAM), log-Mel and Constant Q Transform (CQT) for front-end feature extraction. For back-end classification, we present a re-trained model with a multi-kernel CDNN-based architecture for the pre-trained process and a DNN-based network for the post-trained process. Our obtained results over DCASE 2016 dataset show a significant improvement, increasing by nearly 8% compared to DCASE baseline of 77.2%.

AB - This paper proposes a deep learning framework applied for Acoustic Scene Classification (ASC), which identifies recording location. In general, we apply three types of spectrograms: Gammatone (GAM), log-Mel and Constant Q Transform (CQT) for front-end feature extraction. For back-end classification, we present a re-trained model with a multi-kernel CDNN-based architecture for the pre-trained process and a DNN-based network for the post-trained process. Our obtained results over DCASE 2016 dataset show a significant improvement, increasing by nearly 8% compared to DCASE baseline of 77.2%.

U2 - 10.1109/RIVF48685.2020.9140774

DO - 10.1109/RIVF48685.2020.9140774

M3 - Conference Proceedings with Oral Presentation

SN - 978-1-7281-5377-3 3

SP - 1

EP - 5

BT - RIVF International Conference on Computing and Communication Technologies (RIVF) 2020

T2 - 2020 RIVF International Conference on Computing and Communication Technologies (RIVF)

Y2 - 14 October 2020 through 15 October 2020

ER -

A Re-trained Model Based On Multi-kernel Convolutional Neural Network for Acoustic Scene Classification

Abstract

Konferenz

Research Field

Zugriff auf Dokument

Fingerprint

Diese Publikation zitieren