On the application of SEGAN for the attenuation of the ego-noise in the speech sound source localization problem

In this paper, we present some preliminary results using the Speech Enhancement Generative Adversarial Network (SEGAN) for the attenuation of the ego-noise in the speech source localization problem embedded in unmanned aerial vehicles (UAV). This task is of great interest in UAV search and rescue sc...

Full description

Saved in:
Bibliographic Details
Published in2019 Workshop on Communication Networks and Power Systems (WCNPS) pp. 1 - 4
Main Authors Spadini, Tito, Imai Aldeia, Guilherme Seidyo, Barreto, Guilherme, Alves, Kaleb, Ferreira, Henrique, Suyama, Ricardo, Nose-Filho, Kenji
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.10.2019
Subjects
Attenuation
Autonomous aerial vehicles
direction of arrival
Direction-of-arrival estimation
ego-noise
Location awareness
Real-time systems
SEGAN
Signal processing
Speech enhancement
speech source locatization
Time-frequency analysis
Training
UAV
Wiener filters
Online AccessGet full text

Cover

More Information
Summary:In this paper, we present some preliminary results using the Speech Enhancement Generative Adversarial Network (SEGAN) for the attenuation of the ego-noise in the speech source localization problem embedded in unmanned aerial vehicles (UAV). This task is of great interest in UAV search and rescue scenarios. The primary motivation of using the SEGAN is that it seems to preserve the waveform of the speech signal, which is essential for time-based direction of arrival (TDOA) algorithms. Although preliminary, the obtained results open an excellent perspective for its usage in this problem and despite its computational burden in the training stage, once the SEGAN is trained, it can be implemented for working in real-time scenarios.
DOI:10.1109/WCNPS.2019.8896308