Audio-based detection of malfunctioning machines using deep convolutional autoencoders

Thoidis I, Giouvanakis M, Papanikolaou G. Audio-based detection of malfunctioning machines using deep convolutional autoencoders. In: AES Convention 148. Audio Engineering Society; 2020. Paper 10330. Available from: https://aes.org/publications/elibrary-page/?id=20747

@inproceedings{Thoidis2020_20747,
  author = {Thoidis, Iordanis and Giouvanakis, Marios and Papanikolaou, George},
  title = {{Audio-based detection of malfunctioning machines using deep convolutional autoencoders}},
  booktitle = {AES Convention 148},
  note = {Paper 10330},
  year = {2020},
  month = may,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=20747}
}

TY  - CPAPER
TI  - Audio-based detection of malfunctioning machines using deep convolutional autoencoders
AU  - Thoidis, Iordanis
AU  - Giouvanakis, Marios
AU  - Papanikolaou, George
T2  - AES Convention 148
M1  - Paper 10330
PY  - 2020
DA  - 2020/05/06
UR  - https://aes.org/publications/elibrary-page/?id=20747
PB  - Audio Engineering Society
LA  - en
AB  - In this paper, we develop a modular deep convolutional autoencoder with a dense bottleneck structure to perform the task of unsupervised anomaly detection in machine operating sounds. The proposed model consists of multiple sub-networks with identical encoder-decoder structures, trained to learn a mapping function between different mel-scaled frequency bands. Experiments were conducted on the recently introduced MIMII (Malfunctioning Industrial Machine Inspection and Investigation) open benchmark dataset. Experimental results demonstrate that the proposed model yields improved fault detection performance in terms of the Area Under Curve (AUC) metric compared to the baseline approach.
ER  -