An Environmental Sound Classification Algorithm Based on Multiscale Channel Feature Fusion

Wen Zhao, Helong Wang, Yizi Chen, Xuesong Pan, Kaiyue Zhang, ZhenXiang Bai

2023

Abstract

In recent years, the automatic classification of urban environmental sounds has emerged as a pivotal task in the urban informationization process. Despite the immense potential of environmental sound classification, the accuracy and efficiency of automated classification often fall short of expectations. This paper proposes an environmental sound classification algorithm that integrates multiscale channel feature fusion, aiming to significantly reduce computational complexity while improving classification accuracy. The proposed algorithm comprises two modules: a multiscale channel feature fusion module and a selective feature fusion module, which dynamically merge temporal and frequency domain features. Finally, a series of ablation experiments are conducted and compared with other mainstream algorithms in environmental sound classification, revealing that the proposed algorithm possesses smaller parameter count and higher classification accuracy, thus substantiating the effectiveness of the multiscale channel feature fusion algorithm.

Paper Citation

in Harvard Style

Zhao W., Wang H., Chen Y., Pan X., Zhang K. and Bai Z. (2023). An Environmental Sound Classification Algorithm Based on Multiscale Channel Feature Fusion. In Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT; ISBN 978-989-758-677-4, SciTePress, pages 56-60. DOI: 10.5220/0012273800003807

in Bibtex Style

@conference{anit23,
author={Wen Zhao and Helong Wang and Yizi Chen and Xuesong Pan and Kaiyue Zhang and ZhenXiang Bai},
title={An Environmental Sound Classification Algorithm Based on Multiscale Channel Feature Fusion},
booktitle={Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT},
year={2023},
pages={56-60},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012273800003807},
isbn={978-989-758-677-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT
TI - An Environmental Sound Classification Algorithm Based on Multiscale Channel Feature Fusion
SN - 978-989-758-677-4
AU - Zhao W.
AU - Wang H.
AU - Chen Y.
AU - Pan X.
AU - Zhang K.
AU - Bai Z.
PY - 2023
SP - 56
EP - 60
DO - 10.5220/0012273800003807
PB - SciTePress