Unsupervised Anomaly Detection and Localization of Machine Audio: A GAN-based Approach
This addresses the problem of automatic machine anomaly detection for industrial monitoring, though it is incremental as it builds on existing GAN and autoencoder techniques.
The paper tackles unsupervised anomaly detection and localization in machine audio by proposing AEGAN-AD, a GAN-based method that achieves state-of-the-art results on five machine types in the DCASE 2022 Challenge TASK 2 dataset.
Automatic detection of machine anomaly remains challenging for machine learning. We believe the capability of generative adversarial network (GAN) suits the need of machine audio anomaly detection, yet rarely has this been investigated by previous work. In this paper, we propose AEGAN-AD, a totally unsupervised approach in which the generator (also an autoencoder) is trained to reconstruct input spectrograms. It is pointed out that the denoising nature of reconstruction deprecates its capacity. Thus, the discriminator is redesigned to aid the generator during both training stage and detection stage. The performance of AEGAN-AD on the dataset of DCASE 2022 Challenge TASK 2 demonstrates the state-of-the-art result on five machine types. A novel anomaly localization method is also investigated. Source code available at: www.github.com/jianganbai/AEGAN-AD