Towards cross-modal pre-training and learning tempo-spatial characteristics for audio recognition with convolutional and recurrent neural networks

Towards cross-modal pre-training and learning tempo-spatial characteristics for audio recognition with convolutional and recurrent neural networks

Abstract In this paper, we investigate the performance of two deep learning paradigms for the audio-based tasks of acoustic scene, environmental sound and domestic activity classification. In particular, a convolutional recurrent neural network (CRNN) and pre-trained convolutional neural networks (C...

Full description

Bibliographic Details
Main Authors: Shahin Amiriparian, Maurice Gerczuk, Sandra Ottl, Lukas Stappen, Alice Baird, Lukas Koebe, Björn Schuller
Format: Article
Language:English
Published: SpringerOpen 2020-12-01
Series:EURASIP Journal on Audio, Speech, and Music Processing
Subjects:
Domestic activity classification
Deep learning
Convolutional recurrent neural networks
Deep spectrum
Decision-level fusion
Online Access:https://doi.org/10.1186/s13636-020-00186-0

Internet

https://doi.org/10.1186/s13636-020-00186-0

Similar Items