Deep neural networks for audio scene recognition

Yohan Petetin (1) , Cyrille Laroche (1) , Aurelien Mayoue (1)

1 LADIS (CEA, LIST) - Laboratoire d'analyse des données et d'intelligence des systèmes (CEA, LIST)

These last years, artificial neural networks (ANN) have known a renewed interest since efficient training procedures have emerged to learn the so called deep neural networks (DNN), i.e. ANN with at least two hidden layers. In the same time, the computational auditory scene recognition (CASR) problem which consists in estimating the environment around a device from the received audio signal has been investigated. Most of works which deal with the CASR problem have tried to ind well-adapted features for this problem. However, these features are generally combined with a classical classi-ier. In this paper, we introduce DNN in the CASR ield and we show that such networks can provide promising results and perform better than standard classiiers when the same features are used.

Mots clés: Deep neural network; neural network; deep learnin; audio scene recognition; audio signal processing; belief networks; machine learning; artificial intelligence; neural nets; signal classification; Training; Mel frequency cepstral coefficient; computational auditory scene recognition problem; artificial neural networks; ANN; training procedures; CASR problem; deep belief networks; Index Terms-Deep neural networks; deep beliefs net- works

Domaines: Réseau de neurones [cs.NE]; Intelligence artificielle [cs.AI]

Yohan Petetin, Cyrille Laroche, Aurelien Mayoue. Deep neural networks for audio scene recognition. 2015 23rd European Signal Processing Conference (EUSIPCO), Aug 2015, Nice, France. pp.7362358, ⟨10.1109/EUSIPCO.2015.7362358⟩. ⟨hal-01888746⟩