Text-independent compressed domain speaker verification for digital communication networks call monitoring

In this paper we present a text-independent automatic speaker verification system that works in the compressed domain using GSM AMR coded speech. While traditional approaches process LPC-based cepstral coefficients extracted from LPC related bitstream coefficients, our objective is to study the feasibility of a system that directly processes the raw bitstream transmitted over a digital communication network. In a textindependent closet-set task, with a database of 100 speakers, the proposed system achieves an EER equal to 5.93%, 4.41% and 3.80% for 10, 20 and 30 second long test speech segments respectively.

• Type: Journal
• Year: 2009
• Where: ICMCS
• Authors: Matteo Petracca, Antonio Servetti, Juan Carlos De Martin
• Keywords: ICMCS 2009, Multimedia, Text-independent Automatic Speaker, Textindependent Closet-set Task, Traditional Approaches Process