TF-SepNet: An Efficient 1D Kernel Design in CNNs for Low-Complexity Acoustic Scene Classification

TL;DR

TF-SepNet introduces a novel CNN architecture for acoustic scene classification that uses separate 1D kernels along time and frequency, achieving higher efficiency and better performance than traditional 2D kernel-based models.

Contribution

The paper proposes TF-SepNet, a new CNN design with separate 1D kernels for time and frequency, enhancing efficiency and effectiveness in ASC tasks.

Findings

Outperforms state-of-the-art models with consecutive kernels

Uses less computational resources due to 1D kernel design

Achieves larger effective receptive field for better feature capture

Abstract

Recent studies focus on developing efficient systems for acoustic scene classification (ASC) using convolutional neural networks (CNNs), which typically consist of consecutive kernels. This paper highlights the benefits of using separate kernels as a more powerful and efficient design approach in ASC tasks. Inspired by the time-frequency nature of audio signals, we propose TF-SepNet, a CNN architecture that separates the feature processing along the time and frequency dimensions. Features resulted from the separate paths are then merged by channels and directly forwarded to the classifier. Instead of the conventional two dimensional (2D) kernel, TF-SepNet incorporates one dimensional (1D) kernels to reduce the computational costs. Experiments have been conducted using the TAU Urban Acoustic Scene 2022 Mobile development dataset. The results show that TF-SepNet outperforms similar state-of-the-arts that use consecutive kernels. A further investigation reveals that the separate kernels lead to a larger effective receptive field (ERF), which enables TF-SepNet to capture more time-frequency features.