Robust Audio Event Recognition with 1-Max Pooling Convolutional Neural Networks

TL;DR

This paper introduces a simple yet effective CNN architecture with 1-max pooling and variable filter sizes for robust audio event recognition, achieving state-of-the-art accuracy with fewer layers.

Contribution

The paper proposes a novel, shallow CNN architecture with 1-max pooling and variable filters, outperforming deeper models in audio event recognition tasks.

Findings

Achieved 76.3% relative error reduction over previous models.

Outperformed deep CNN architectures by up to 4.5% in accuracy.

Demonstrated robustness and efficiency with a three-layer network.

Abstract

We present in this paper a simple, yet efficient convolutional neural network (CNN) architecture for robust audio event recognition. Opposing to deep CNN architectures with multiple convolutional and pooling layers topped up with multiple fully connected layers, the proposed network consists of only three layers: convolutional, pooling, and softmax layer. Two further features distinguish it from the deep architectures that have been proposed for the task: varying-size convolutional filters at the convolutional layer and 1-max pooling scheme at the pooling layer. In intuition, the network tends to select the most discriminative features from the whole audio signals for recognition. Our proposed CNN not only shows state-of-the-art performance on the standard task of robust audio event recognition but also outperforms other deep architectures up to 4.5% in terms of recognition accuracy, which is equivalent to 76.3% relative error reduction.