Saved in:
Bibliographic Details
Main Authors: Ye, ShuQi, Tian, Yuan
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2405.03567
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914785904295936
author Ye, ShuQi
Tian, Yuan
author_facet Ye, ShuQi
Tian, Yuan
contents Acoustic scene classification (ASC) is highly important in the real world. Recently, deep learning-based methods have been widely employed for acoustic scene classification. However, these methods are currently not lightweight enough as well as their performance is not satisfactory. To solve these problems, we propose a deep space separable distillation network. Firstly, the network performs high-low frequency decomposition on the log-mel spectrogram, significantly reducing computational complexity while maintaining model performance. Secondly, we specially design three lightweight operators for ASC, including Separable Convolution (SC), Orthonormal Separable Convolution (OSC), and Separable Partial Convolution (SPC). These operators exhibit highly efficient feature extraction capabilities in acoustic scene classification tasks. The experimental results demonstrate that the proposed method achieves a performance gain of 9.8% compared to the currently popular deep learning methods, while also having smaller parameter count and computational complexity.
format Preprint
id arxiv_https___arxiv_org_abs_2405_03567
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Deep Space Separable Distillation for Lightweight Acoustic Scene Classification
Ye, ShuQi
Tian, Yuan
Sound
Artificial Intelligence
Audio and Speech Processing
Acoustic scene classification (ASC) is highly important in the real world. Recently, deep learning-based methods have been widely employed for acoustic scene classification. However, these methods are currently not lightweight enough as well as their performance is not satisfactory. To solve these problems, we propose a deep space separable distillation network. Firstly, the network performs high-low frequency decomposition on the log-mel spectrogram, significantly reducing computational complexity while maintaining model performance. Secondly, we specially design three lightweight operators for ASC, including Separable Convolution (SC), Orthonormal Separable Convolution (OSC), and Separable Partial Convolution (SPC). These operators exhibit highly efficient feature extraction capabilities in acoustic scene classification tasks. The experimental results demonstrate that the proposed method achieves a performance gain of 9.8% compared to the currently popular deep learning methods, while also having smaller parameter count and computational complexity.
title Deep Space Separable Distillation for Lightweight Acoustic Scene Classification
topic Sound
Artificial Intelligence
Audio and Speech Processing
url https://arxiv.org/abs/2405.03567