Year: 2018
Type: conference-paper
Abstract: We consider the problem of learning to localize a speech source using a humanoid robot equipped with a binaural hearing system. We aim to map binaural audio features into the relative angle between the robot's head direction and the target source direction based on a sensorimotor training framework. To this end, we make the following contributions: (i) a procedure to automatically collect and label audio and motor data for sensorimotor training; (ii) the use of a convolutional neural network (CNN) trained with white noise signal and ground truth relative source direction. Experimental evaluation with speech signals shows that the CNN can localize the speech source even without an explicit algorithm for dealing with missing spectral features.
Source: HAL
Language: English
Topic: Speech and Audio Processing
Subfield: Signal Processing
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2903315366
