eduzhai > Applied Sciences > Engineering >

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

  • Save

... pages left unread,continue reading

Document pages: 5 pages

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

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×