#################################################
#################################################
This project is concerned with multi-speaker voice activity detection and source enumeration in wireless acoustic sensor networks (WASN). The performance of the traditional technique based on multiplicative nonnegative ICA decreases as the number of speaker increases. The proposed technique first clusters the nodes that observe a single speaker as dominant source and then estimates the voice activity of each speaker by introducing a block-sparsity penalizing term in the unmixing problem. The results are validated with WASN speech dataset that has been generated within the EU FET-Open Project HANDiCAMS (GA no. 323944).
###########
###########
In case of questions, suggestions, problems etc. please send an email.
Tanuj Hasija: [email protected]
Martin Gölz [email protected]
##############
##############
[1] T. Hasija, M. Gölz, M. Muma, P. J. Schreier and A. M. Zoubir, "Source Enumeration and Robust Voice ActivityDetection in Wireless Acoustic Sensor Networks," Proc. Asilomar Conf. Signals Syst. Computers, Pacific Grove, CA, USA, November 2019.