Anomalous Sound Detection using unsupervised and semi-supervised autoencoders and gammatone audio representation

TL;DR

This paper introduces a novel unsupervised and semi-supervised autoencoder framework utilizing Gammatone audio representations for effective anomalous sound detection, particularly in industrial settings, outperforming existing baselines.

Contribution

It presents a new framework combining convolutional autoencoders and Gammatone features for improved unsupervised anomalous sound detection.

Findings

Significant performance improvement over baseline methods

Effective early detection of machine malfunctions

Applicable to real-world industrial scenarios

Abstract

Anomalous sound detection (ASD) is, nowadays, one of the topical subjects in machine listening discipline. Unsupervised detection is attracting a lot of interest due to its immediate applicability in many fields. For example, related to industrial processes, the early detection of malfunctions or damage in machines can mean great savings and an improvement in the efficiency of industrial processes. This problem can be solved with an unsupervised ASD solution since industrial machines will not be damaged simply by having this audio data in the training stage. This paper proposes a novel framework based on convolutional autoencoders (both unsupervised and semi-supervised) and a Gammatone-based representation of the audio. The results obtained by these architectures substantially exceed the results presented as a baseline.