Receptive-field-regularized CNN variants for acoustic scene classification

TL;DR

This paper investigates how restricting the receptive field of CNNs affects acoustic scene classification, introduces Frequency Aware CNNs to recover lost frequency information, and demonstrates improved performance on DCASE 2019 tasks.

Contribution

It systematically studies RF configurations in CNNs for audio tasks and proposes Frequency Aware CNNs to enhance performance when RF is restricted.

Findings

Restricted RF impacts CNN performance and robustness.

Frequency Aware CNNs improve classification accuracy.

Several submissions achieved top results in DCASE 2019 Challenge.

Abstract

Acoustic scene classification and related tasks have been dominated by Convolutional Neural Networks (CNNs). Top-performing CNNs use mainly audio spectograms as input and borrow their architectural design primarily from computer vision. A recent study has shown that restricting the receptive field (RF) of CNNs in appropriate ways is crucial for their performance, robustness and generalization in audio tasks. One side effect of restricting the RF of CNNs is that more frequency information is lost. In this paper, we perform a systematic investigation of different RF configuration for various CNN architectures on the DCASE 2019 Task 1.A dataset. Second, we introduce Frequency Aware CNNs to compensate for the lack of frequency information caused by the restricted RF, and experimentally determine if and in what RF ranges they yield additional improvement. The result of these investigations are several well-performing submissions to different tasks in the DCASE 2019 Challenge.