Deep Dense and Convolutional Autoencoders for Machine Acoustic Anomaly Detection - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2021

Deep Dense and Convolutional Autoencoders for Machine Acoustic Anomaly Detection

Gabriel Coelho
  • Function : Author
  • PersonId : 1105401
Pedro Pereira
  • Function : Author
  • PersonId : 1105402
Luis Matos
  • Function : Author
  • PersonId : 1105403
André Ferreira
  • Function : Author
  • PersonId : 1105404
Paulo Cortez
  • Function : Author
  • PersonId : 1004076
André Pilastri
  • Function : Author
  • PersonId : 1105405

Abstract

Recently, there have been advances in using unsupervised learning methods for Acoustic Anomaly Detection (AAD). In this paper, we propose an improved version of two deep AutoEncoders (AE) for unsupervised AAD for six types of working machines, namely Dense and Convolutional AEs. A large set of computational experiments was held, showing that the two proposed deep autoencoders, when combined with a mel-spectrogram sound preprocessing, are quite competitive and outperform a recently proposed AE baseline. Overall, a high-quality class discrimination level was achieved, ranging from 72% to 92%.
Embargoed file
Embargoed file
0 6 23
Year Month Jours
Avant la publication

Dates and versions

hal-03287664 , version 1 (15-07-2021)

Licence

Attribution

Identifiers

Cite

Gabriel Coelho, Pedro Pereira, Luis Matos, Alexandrine Ribeiro, Eduardo C. Nunes, et al.. Deep Dense and Convolutional Autoencoders for Machine Acoustic Anomaly Detection. 17th IFIP International Conference on Artificial Intelligence Applications and Innovations (AIAI), Jun 2021, Hersonissos, Crete, Greece. pp.337-348, ⟨10.1007/978-3-030-79150-6_27⟩. ⟨hal-03287664⟩
31 View
29 Download

Altmetric

Share

Gmail Facebook Twitter LinkedIn More