CSR: SPEAKER RECOGNITION FROM COMPRESSED VOIP PACKET STREAM (WedPmPO1)

Author(s) : Charu Aggarwal (IBM, United States of America) David Olshefski (IBM, United States of America) Saha Debanjan (IBM, United States of America) Zon-Yin Shae (IBM, United States of America) Philip Yu (IBM, United States of America)

Abstract : VoIP applications require the ability to identify speakers in real time. This paper presents Compressed Speaker Recognition (CSR), an innovative approach to perform speaker recognition directly from the compressed voice packets. CSR performs online speaker recognition from live packet streams of compressed voice packets by performing fast clustering over a defined subset of the features available in each compressed voice packet. Our experimental results show that CSR is highly scalable and accurate across a brad range of speakers.

Menu