Speaker localization and separation using incremental distributed expectation-maximization

Add to your list(s) Download to your calendar using vCal

• Yuval Dorfan, Bar-Ilan University, Ramat-Gan, Israel
• Monday 17 October 2016, 13:00-14:00
• R503, EEE Department.

If you have a question about this talk, please contact Alastair Moore.

Abstract

A network of microphone pairs is utilized for the joint task of localizing and separating multiple concurrent speakers. The recently presented incremental distributed expectation-maximization (IDEM) is addressing the first task, namely detection and localization. Here this algorithm is extended to address the second task, namely blindly separating the speech sources. It is shown that the proposed algorithm, denoted distributed algorithm for localization and separation (DALAS), is capable of separating speakers in reverberant enclosure without a priori information on their number and locations. In the first stage of the proposed algorithm, the IDEM algorithm is applied for blindly detecting the active sources and to estimate their locations. In the second stage, the location estimates are utilized for selecting the most useful node of microphones for the subsequent separation stage. Separation is finally obtained by utilizing a byproduct of the IDEM algorithm, namely its hidden variables to construct masks for each source in the relevant node. The estimated source locations also enable coherent averaging of the node’s signals received by different microphones. Encouraging simulation results demonstrate the potential of the algorithm to blindly separate sources even in close proximity.

Biography

Yuval Dorfan received his B.Sc. degree (summa cum laude) from Ben-Gurion University, Beer-Sheva, Israel in 1998, his M.Sc. degree (magna cum laude) from the Technion Israel Institute of Technology, Haifa, Israel in 2000 both in electrical engineering and his MBA degree from the Interdisciplinary Center, Herzliya, Israel in 2007. He is currently pursuing the Ph.D. degree at the Engineering Faculty in Bar-Ilan University, Ramat-Gan, Israel. He has twenty years of research and development experience in various fields of signal processing and communication industries. Mr. Dorfan has been already published a few patents conference papers and journal papers. His research interests include distributed sensor networks, distributed acoustic source localization, blind source separation and distributed speaker tracking.

This talk is part of the COMMSP Seminar series.

This talk is included in these lists:

• COMMSP & CP list
• COMMSP Seminar
• R503, EEE Department
• SAP Talks

Note that ex-directory lists are not shown.